The initial decision of a farmer in a developing country to adopt a new agricultural
technology is related to the decisions of the people in his social network – his friends and
family. At first, the more that others have adopted the technology, the more likely he is to do
the same; but when the number of adopters has become much bigger, his probability of
adopting the technology actually falls. For the better-informed farmers, however, the
adoption choices of others are less significant in their own decisions.
These are the central findings of new research by Oriana Bandiera and Imran Rasul,
published in the October 2006 Economic Journal. While their analysis focuses on the
specific experience of farmers in rural Mozambique, the research has far broader
applicability: similar patterns of adoption decisions within social networks may occur in all
sorts of economic environments in which a new technology is introduced, information is a
key barrier to adoption and individuals learn about the new technology from others.
The adoption of new agricultural technologies is an important route out of poverty for many
in the developing world. Yet agricultural innovations are often adopted slowly and some
aspects of the adoption process remain poorly understood. Bandiera and Rasul analyse the
decision to adopt a new crop – sunflower – by farmers in the Zambezia province of
Northern Mozambique.
While intuition suggests that farmers should be more likely to adopt when others in their
social network do so, theories of social learning indicate that the relationship is ambiguous.
On the one hand, the benefit of adopting is higher when there are many adopters in the
network because of the information they provide. On the other hand, having many adopters
in the social network increases incentives to free ride on the knowledge accumulated by
others. If such considerations prevail, a farmers' propensity to adopt decreases as the
number of adopters among his network increase.
The researchers measure the information on sunflower cultivation available to each farmer
from his social network as the number of adopters among his self-reported network of
family and friends. They then estimate farmers' propensity to adopt sunflower as a function
of the number of adopters among their family and friends.
They find that the relationship between the probability of a given farmer adopting the new
crop, and the number of his friends and family that do so, is an inverted U-shape. In other
words, the probability of adopting increases at first but then falls as the number of adopters
in the network becomes very large.
The estimates indicate that the impact of social networks on adoption is large compared
with the effect of individual determinants, such as the age and the vulnerability of the
farmer. Moving from having no adopters in the network to having between 1 and 5
increases the propensity to adopt by 0.27, while having between 6 and 10 increases the
propensity by 0.58.
In other words, a farmer is more likely to adopt than not if he has between 6 and 10
adopters among his family and friends. But having more than 10 adopters increases the
propensity to adopt by only .30, relative to having no adopters in the network.
The study also finds that the effect of the number of adopters among the network on each
farmer’s propensity to adopt depends on the characteristics of the farmer himself. In
particular, more informed farmers are less sensitive to the adoption choices of their
network.
ENDS
Notes for editors: ‘Social Networks and Technology Adoption in Northern Mozambique’ by
Oriana Bandiera and Imran Rasul is published in the October 2006 issue of the Economic
Journal.
Oriana Bandiera is at the London School of Economics. Imran Rasul is at University
College London.
For further information: contact Imran Rasul on 0207-679-5853 (email:
i.rasul@ucl.ac.uk); or Romesh Vaitilingam on 07768-661095 (email:
romesh@compuserve.com).