Markets in developing countries are poorly integrated, especially in rural areas. An important consequence of this is that rural markets see tremendous seasonal price variation - in Western Kenya for example, we observed an approximately 40% increase in prices between the peak and trough of the agricultural season, while Burke (2014) reports increases of 87% in another part of Western Kenya. Farmers are universally aware of this, and many worry about how they will cope with food price increases during the "hungry season" just prior to the harvest. Despite being aware of the impending surge in prices, many farmers find it difficult to save their harvest, and a significant fraction sell maize immediately after harvest at rock-bottom prices. In addition, several also buy later in the year at much higher prices in order to meet consumption needs.
The main difficulty that farmers report facing in storing maize is the obvious one: people simply lack a good place to store maize. Nearly all farmers in a pilot study we conducted in 2015-16 hold at least some maize after harvest (91%) but the technology is extremely inefficient: most people simply keep the maize in the house on a table or on the floor. Farmers estimate that 25% of the maize they store is spoiled due to moisture or pests and rodents, (indeed, 95% of our respondents report such spoilage as one of the challenges to storing maize). Moreover, when maize sits at home in plain sight, it often falls prey to temptation or to claims by others: 26% of our respondents feel that they consume “too much” when they have bags of maize at home. A parallel problem is that friends or relatives often ask for help, and it is hard to say no when the maize is in plain sight in the house. Over 60% of respondents say that they try to find ways to give out as little maize as possible to these sorts of requests, but many find it difficult to say no.
Our idea is to encourage a cost-effective, scalable, group-based intervention to improve this situation, which we call the group savings and reinvestment account (GSRA). This intervention will harness the existence of ROSCAs (Rotating Savings and Credit Associations, i.e. group savings clubs) in Western Kenya. The main innovation is providing farmers with a simple technology for storing maize: we will provide each ROSCA with insulated air-tight plastic bags, designed specifically for the purpose of storing grain (these are called the Purdue Improved Crop Storage (PICS) bags: https://ag.purdue.edu/ipia/pics/Pages/home.aspx). We will also provide a wooden stand, to keep the maize elevated from the ground (and less susceptible to pests and water damage).
The GSRA intervention may encourage usage through a variety of channels. The key factors here are that the bags and stand are a technological improvement over saving in the home; that the group nature of the ROSCA may help provide some “peer pressure”; that setting aside maize in the GSRA (where the maize will be out of the sight of the farmer) may keep the maize “out of sight and out of mind”; and that the GSRA may encourage the activation of a mental account for input usage. The GSRA may also make complementary investments (such as for fertilizer) more profitable, by increasing the effective price that farmers receive for their output.
This project will be a research project, conducted by Jonathan Robinson (associate professor of economics, UC Santa Cruz and research affiliate at IPA, CEGA, JPAL, NBER, and BREAD), Shilpa Aggarwal (assistant professor of economics, Indian School of Business), and Eilin Francis (graduate student in economics, UC Santa Cruz). Our aim is to evaluate the effect of the GSRA through a randomized controlled trial (RCT). A randomly selected set of ROSCAs will receive the GSRA treatment, while a control group will receive nothing. Since the groups are randomly selected, any post-treatment differences in outcomes will be attributable to the GSRA, rather than to some other factor. The evaluation of this RCT will be conducted in collaboration with Innovations for Poverty Action - Kenya (http://www.poverty-action.org/).
Research Design & Data Collection
At an initial meeting, each ROSCA (including the control group) will be read a script about the benefits of setting maize aside after the harvest, of using inputs generally, and of savings. The basic script will be augmented for the GSRA ROSCAs to encourage members to collectively set aside some portion of their 2016 long rains harvest using the bags and the stand that we will subsidize, and hold it till the time the selling price of maize is more favorable. Each ROSCA will identify a member who will house the stand to store the group's maize till it is sold, individually or collectively. GSRA ROSCAs will also be provided with a ledger book to enable the treasurer to keep track of all deposits and withdrawals of maize by individual members.
We will collect several pieces of data. First, at the first ROSCA meeting, we will collect a baseline survey for a random sample of the ROSCA members, which will focus on savings, expectations about prices, and constraints to storing maize (as well as basic demographic information). Second, we will visit ROSCAs to collect the administrative records on take-up and usage of the accounts. Third, we will conduct endline surveys. In addition to other questions, this survey will focus on total maize storage (to measure crowd out from the GSRA treatment - i.e. if people do use the GSRA, is this just a substitution from savings at home, or is it a real increase in savings?). The endline will also enable us to measure the extent to which farmers were able to take advantage of intertemporal arbitrage opportunities.
Thus far, we have piloted this intervention with 141 ROSCAs and 3,105 respondents. Preliminary take-up results are very encouraging: 96% of GSRA ROSCAs took up the product, and the average ROSCA is storing 260 kilograms of maize, an amount worth about $57 per ROSCA at post-harvest prices and $108 if held until later. At the individual level, we find that about 40% of farmers set aside some maize in the GSRA; the average amount held by these individuals was 32 kilograms, which is equivalent to roughly 5% of reported harvest output (609 kg). Five percent of farm income is a relatively substantial sum, especially given that much of maize people typically hold is for consumption but the maize held in the GSRA is intended to be illiquid.
In a survey, we find evidence that the GSRA had real effects on savings. First, we find that GSRA ROSCAs held more total maize than control ROSCAs - it appears that most, or all, of the maize held in the GSRA represented new savings (rather than solely a reshuffling of money from home to the GSRA). Second, people reported being happy with the GSRA: 91% report that the GSRA enabled them to save more. There are a mix of reasons for this: people report that they saved more directly since the bags are less prone to pests/spoilage; however, other important reasons included that people gave out less maize when it was kept outside the home, and that people "over-consumed" less.
A grant from IDEO would allow us to explore these issues much more deeply. While we view this preliminary take-up as encouraging, there are several reasons to suspect this may be a lower-bound on true usage. First, we conducted this pilot over summer 2015, in preparation for the 2015 long rains harvest (which was in August). However, due to tight timing, many of the visits were conducted just before, or in some cases just after the harvest. Effects may be bigger if farmers anticipate this product being available earlier. Second, many ROSCAs (80% of the GSRA group) reported that that their members contributed less than their desired amount. The reasons cited for this were that the harvest was less than expected (possibly suggesting higher take-up in other seasons), and that the intervention started too late. Third, the pilot was limited in that we only provided 4 bags for ROSCAs in this round. Several ROSCAs reported this as a capacity constraint. If funded, we will provide more bags to ROSCAs in future seasons. Fourth, this is a new product and people may want to “try it out” with a smaller sum in the first year.