When it comes to spreading information through a social network, researchers studying participation in a microfinance project in India find it really is about who you know.
When the representatives from Bharatha Swamukti Samsthe, a microfinance institution in India, want to get the word out about a new loan program, they a handful of people – teachers, shopkeepers and heads of local savings groups – and ask them to spread the news.
These village leaders, it’s assumed, will be most effective in informing their friends and neighbors about the program and getting people to participate.
But does it really work that way?
New research by Stanford and MIT economists looks at who the best people are to to spread information through a social network and whether their influence affects participation.
The researchers, Matthew Jackson and Arun Chandrasekhar of Stanford and Esther Duflo and Abhijit Banerjee of MIT, developed a new measure of social influence they call "diffusion centrality." The model incorporates not only how many friends a person has but how well connected a person is to other people who are well connected.
Looking at the operations of BSS in 43 villages outside Bangalore, the researchers found that participation in the microfinance program varied from about 7 percent to 44 percent, depending on who was initially told.
"In villages where they contacted very central people, they ended up with high participation, while in villages where they contacted less central people they ended up with low participation," Jackson said.
Further, the researchers found that participation in loan programs increases by about 11 percentage points when well-connected local residents are the first to gain access to the programs.
"This suggests that policymakers may have a lot to gain – or lose – by hitting the right people first," Chandrasekhar said.
Their paper, "The Diffusion of Microfinance," was published this month in the journal Science.
Jackson described the research as a unique opportunity to witness how social network structure affects human behavior.
Six months before the bank went into the villages to offer loans, the researchers collected detailed information about the local residents, everything from who borrows kerosene and rice from whom, to who gives advice to whom, and who goes to temple with whom.
"We were able to see how the network structure affected the diffusion of the loan program through the villages," Jackson said.
Among the findings, the researchers discovered that program participants are seven times more likely to pass along information about the program to other households compared to nonparticipants. However, information passed by nonparticipants still accounts for roughly one-third of eventual participation.
They also found that an informed individual is not more likely to participate if his or her informed friends participate.
"Instead they seem to make relatively independent decisions which can depend on their personal characteristics," Chandrasekhar said.
The researchers said their findings should be helpful to many different organizations that might be interested in diffusing information about a new product or program. They cautioned, however, that each situation has its own nuances.
Peter Dizikes of MIT News contributed to this report.
For more Stanford experts on economics and other topics, visit Stanford Experts.
Matthew Jackson, Economics: jacksonm [a] stanford (p) edu
Arun Chandrasekhar, Economics: arungc [a] stanford (p) edu
Brooke Donald, Stanford News Service: (650) 725-0224, brooke.donald [a] stanford (p) edu