Policy Brief Issue 15 | March 2016 AS GOOD AS THE COMPANY THEY KEEP? IMPROVING FARMERS’ SOCIAL NETWORKS GENDER Authors: Kenneth Leonard and Kathryn Vasilaky INNOVATION LAB KEY MESSAGES The Gender Innovation Lab (GIL) conducts impact • Extension services have a history of being relatively evaluations of development expensive and not always effective. At the same time, interventions in Sub-Saharan Africa, seeking to generate studies show that informal social networks can be evidence on how to close very beneficial in helping increase productivity. the gender gap in earnings, productivity, assets and • In Uganda, we tested the value of informal social networks agency. The GIL team is for women farmers by connecting the least-productive 30% currently working on over to some of the most productive women farmers in their own 50 impact evaluations in 21 countries with the aim of villages. Results show significant gains in productivity indicating that building an evidence base the path to better outcomes is contained within their own community. with lessons for the region. • Women learned the agricultural information at least as The impact objective of GIL is well in a network setting as in a more intensive, formal increasing take-up of effective policies by governments, extension setting. development organizations • On average, the social network intervention was less and the private sector in order to address the underlying costly and more effectively targeted women and the least causes of gender inequality productive farmers than traditional extension services. in Africa, particularly in terms of women’s economic and Social networks affect all of our lives; the people we know influence us every social empowerment. The lab day, including by sharing their knowledge and experience. If something aims to do this by producing happens inside our social network, we are more likely to know about it. On the and delivering a new body of evidence and developing a other hand, if our networks do not extend to many individuals, the less new compelling narrative, geared information we are exposed to. This is especially true for people living in poor towards policymakers, on what and remote areas where there is limited access to other sources of information. works and what does not work in promoting gender equality. For more information visit us at: http://www.worldbank.org/en/programs/africa-gender-innovation-lab Traditionally, approaches to increasing agricultural The researchers used a two-armed randomized productivity have focused on extension programs controlled trial (RCT) to compare a standard that target men because they are often more visible. agricultural training program, which targeted men Often the expectation is that the extension service and women, with a social network intervention that trainees will disseminate new information to other only targeted women. The standard agricultural farmers in their village, but this rarely happens. training program was previously implemented by This setup is particularly costly for women who extension agents and involved bi-weekly meetings are frequently outside the social networks that with participants. This training included male and would allow them to learn about these innovations. female cotton farmers who were invited to participate The need to more effectively reach women with in a series of training sessions during the five critical agricultural information is particularly urgent as we stages of the cotton season including pre-planting, already know that there are large gender gaps in planting, pesticide use, harvesting and marketing. agricultural productivity across Sub-Saharan Africa. For the social network intervention (SNI), the team invited female cotton farmers to a networking session and paired HERE’S WHAT WE DID each woman with another female cotton farmer whom Researchers at the University of Maryland thought they did not already know. The paired women were that linking women who farmed the same crop, given photos of each other and asked to speak to each yet did not know of one another’s experiences, other throughout the cotton-growing season. During might improve the relevance of women’s social the networking session, the paired women identified networks to their farming activities and so allow cultivation issues, chose a collaborative goal, and set them to more effectively spread information that times when they would meet to exchange information in could help them to improve their productivity. the future. The session also involved a learning game, led by two external facilitators, in which incentives were given With support from the World Bank’s Africa Gender for learning and remembering new cotton farming facts.1 Innovation Lab, the researchers decided to test this theory with an experiment involving cotton farmers in Uganda. This was a highly appropriate HERE’S WHAT WE FOUND context for the experiment for three reasons: (1) A key element to the success of the social cotton was only recently reintroduced to certain network intervention was that it allowed regions of Uganda after the civil war and political less productive women to learn from more unrest, so there was an absence of knowledge productive women within their own village. about the best cotton growing techniques; (2) Even though this intervention may not benefit the average productivity was low; and (3) extension top-performing farmers directly it is important services offered for cotton growers were not that they are incorporated in the intervention targeted to women—one suggested reason as for the benefit of the other farmers. Also, being to why a large gap in crop yields between male paired with a worse farmer did not reduce and female-headed households existed. the yields of the top-performing farmers. 1 The same extension training materials and facts were used in the SNI as in the Training arm to teach women about cotton. IMPACT OF TRADITIONAL TRAINING VS. SOCIAL NETWORK INTERVENTION Social Network Intervention 98 kg/acre Traditional Training 67 kg/acre The social network approach was at least By establishing new network ties with fellow cotton as effective at spreading new agricultural farmers whom they did not previously know, and gaining information as formal training. new information, participants were able to increase their agricultural productivity. This was especially true for the Overall, womens’ yields increased by 67 kgs/ least productive female farmers. The impact appears to acre under the Formal Training program and 98 work by improving the suitability of women’s networks kilograms/acre under the SNI, a large increase for learning: after the intervention, participants’ networks from the average yield (180kgs/acre). To test consisted of a higher proportion of female cotton farmers. whether these gains can be attributed to increased It was found that increasing the proportion of women in knowledge that women acquired in training or a woman’s social network by 10 percentage points leads through talking to each other, a quiz of cotton- to an average increase in yield of 26 kilograms per acre. farming knowledge was administered. Results at follow up showed that participants improved their While the impacts were greatest for women, there were scores by between 4 and 5%, which translates to also impacts on men. Villages that participated in the an increase in yields of around 15 kilograms per social network intervention observed significant spillovers acre. Thus, improved knowledge on the points that in productivity to men farmers, though not as great as the farmers were quizzed on account for some those of women who participated in the intervention. In part of the gains. Other potential mechanisms for comparison, the formal training intervention increased the gains from the SNI were the overall interaction yields for men significantly more than it did so for women. with one’s pair, information learned beyond that of the quiz material, and potentially, access to The overall cost of the traditional training program was higher performing farmers outside one’s network. approximately $20 USD per farmer for 354 trained farmers. Each extension trainer worked with between 7 and 30 farmers on a bi-weekly basis. The comparable cost for the On average, the social network intervention SNI was $5 per farmer, 25% of the cost of the traditional was less costly and more effectively targeted training. Furthermore, there was less variation in how women and the least productive farmers the training was provided. Each SNI trainer handled 14 than traditional extension services. farmers and followed the same protocol in each session. OVERALL COST OF TRADITIONAL TRAINING VS. SOCIAL NETWORK INTERVENTION $5 USD / FARMER Social Network The overall cost of Intervention the Social Network Intervention was 25% of the Traditional Training costs. $20 USD / FARMER Traditional Training Thus, the social network intervention was significantly less expensive than the traditional training intervention, yet it achieved at least the same impact on yields on average and demonstrated its highest impacts for the least productive females. CONCLUSION By exploiting the power of social ties, social network interventions offer a lower-cost alternative to traditional agricultural training programs and can be particularly effective at improving the productivity of women. The results of the study featured in this brief are particularly relevant to policymakers in Sub- Saharan Africa, where productivity differentials still exist between males and females, and women are less frequently targeted for training. FOR MORE INFORMATION, For more information on this study see the paper: Vasilaky, Kathryn PLEASE CONTACT and Kenneth L. Leonard. 2015. “As Good as the Networks They Keep? : Improving Farmers’ Social Networks via Randomized Information Exchange Markus Goldstein mgoldstein@worldbank.org in Rural Uganda.” Rachel Coleman rcoleman1@worldbank.org 1818 H. St NW Washington, DC 20433 USA