Naïve learning as a coordination device in social networks

Abstract

I propose a framework describing how naïve learning in networks may determine coordination outcomes of product adoption. Individuals receive initial signals regarding the value of the product, communicate afterwards, and make adoption decisions based on that. In the framework of DeGroot's Naïve Learning, the model suggests that as beliefs converge, the result will converge to a unique cutoff equilibrium, similar to a global game. I then describe how adoption rates and social welfare depend on network structures by showing that the variance of the unit eigenvector centrality of the listening matrix, which represents inequality in network positions, is a sufficient statistic for adoption in equilibrium. More adoption is expected with high inequality in network positions if the value of the product to be adopted is low, and vice versa. The relationship between social welfare and inequality in network positions aligns with that of adoption and inequality in network positions, except in cases of very low product value, where increased adoption may reduce overall social welfare.