Last week we blogged about the puzzle of low adoption of mobile banking accounts in Bangladesh. When we ask, people say they’re interested in mobile banking accounts. So why don’t they adopt the new technology? And why are rates of adoption so high in some places, like Kenya, and so low in Bangladesh?
This week we zoom in on one particular aspect of the adoption puzzle, and discuss the possibility that there are social phenomena (“peer effects” as they are known in the social science literature) that help to explain low adoption rates in some contexts and high adoption rates in others. An idea originally proposed by Nobel Laureate Thomas Schelling and more recently popularized by writers like Malcolm Gladwell is that peer effects can generate so-called “tipping points” in aggregate behavior.
A typical example involves how people seat themselves when entering an empty auditorium. If the first person to enter sits near the front, the second and third people to enter might do the same. Preferring to do what others before them have done, new arrivals might continue to fill the auditorium towards the front. However, if the first person to enter sits near the back of the auditorium, a different outcome might emerge in which the audience clusters at the back of the auditorium. When people’s choices depend on the choices of those around them, in many economic models, multiple outcomes or equilibria can emerge.
To add a little more complexity, my friend’s choices might influence my choices in one or both of two ways. First, I might learn something about the existence or usefulness of a given product that might cause me to make the same choice (social learning). Second, I might directly care about the choices that my friend makes – and make the same choice because I know she made it. It’s this second type of peer influence that can really generate tipping points that matter across a range of economic contexts.
A new experimental paperby Leonardo Bursztyn, Florian Ederer, Bruno Ferman and Noam Yuchtman looks at a new financial product in Brazil and tries to quantify how important each of these types of peer influence are in explaining why people decide to invest in the new product or not. The paper, which is forthcoming in the journal Econometrica, uses an innovative randomization among pairs to tease apart peer effects due to social learning from the type of peer influence which can generate tipping points, which they term “social utility.” In this second model of peer influence, individuals value the new product more purely because their peers or social contacts have adopted it.
What they find is that both types of peer effects matter – a lot. Investors who are not informed of their peers’ choices adopt at a rate of 42 percent. Investors who only learn that their peers would value the product enough to buy adopt at a rate of 71 percent. Finally, investors who learn that their peers would value the product enough to buy and that their peers actually purchased the asset adopt at a rate of 93 percent.
Similarly, in our study, the overall utility of mobile banking might vary with the adoption choices of family and friends. Not only do individuals learn about the potential usefulness of the product from the adoption choices of their family and friends, but in the case of mobile banking, there is a very clear network externality from the remittance feature of the product. It’s more useful to have an account if your social contacts are on the same network, and can send and receive money via their accounts.
Although it will be more challenging in our context, we’re hoping to implement a similar experimental design to learn about whether social utility or network externalities might have a significant role to play in explaining persistently low rates of adoption of mobile banking accounts in Bangladesh.
We’re hoping to learn something that has real implications for its marketing – if it’s fundamentally a good product, but no one has adopted because their friends haven’t, then there’s a potential role for a “big push” marketing campaign. We are collaborating with bKash and with the central bank of Bangladesh in the study, and so we hope that the lessons we learn about peer or network externalities from our experiment will inform policies in the future.
Photo by Jean Lee: bKash agent outside of his shop in Bangladesh