Social network externalities and price dispersion in online markets

Ample empirical studies in the e-commerce literature have documented that the price dispersion in online markets is 1) as large as that in offline markets, 2) persistent across time, and 3) only partially explained by observed eretailers’ attributes. Buying on the internet market is risky to consumers. First of all, consumers and the products they purchase are separated in time. There is a delay in time between the time consumers pay and the time they receive the orders. Second, consumers and the products they purchase are separated in space. Consumers cannot physically touch or examine the products at the point of purchase. As such, online markets involve an adoption process based on the interaction of consumers’ experiences in the form of references, recommendations, word of mouth, etc. The social network externalities introduced by the interaction of consumer’s experiences reduces the risk of seller choice and allows some sellers to charge higher prices for even homogeneous products. This research aims to study online market price dispersion from the social network externalities perspective. Our model posits that consumers are risk averse and assess the risk of having a satisfactory transaction from a seller based on the two dimensions of the seller’s social network externalities: quantity externality (i.e., the size of the seller’s social network) and quality externality (i.e., the satisfactory transaction probability of the seller’s social network). We further investigate the moderating effect of product value for consumers on the impact of social network externality on online market price dispersion. Our model yields several important propositions which we empirically test using data sets collected from eBay. We found that 1) both quantity externality and quality externality of social network are salient in driving online price dispersion, and 2) the salience of social network externality is stronger for purchase behavior in higher value product categories.