Econometrics of Network Formation

Abstract

This chapter surveys econometric network formation models. Its goal is to acquaint the readers, in a self-contained manner, with a number of network formation models used in the graph theory, statistics, sociology, and econometrics literature, with a view to how well they map to real-world data, the sorts of economic microfoundations they implicitly assume, and their econometric properties. A major difficulty in the study of network formation is that the researcher typically has a data set consisting of a single network observed in a single period. Key questions include whether a researcher can identify, and develop consistent estimators of, the parameters driving network formation when observing a single large network. Estimation involves a number of challenges, including but not limited to the degree of correlation between linking decisions, the ability to reproduce realistic patterns of network structure, concerns about multiple equilibria, and missing data.

The formation of physician patient sharing networks in medicare: Exploring the effect of hospital affiliation

“…One hurdle in accomplishing this aim has been that establishing a credible case for causality in the presence of unobservable physician specific characteristics that may play a role in physician patient sharing (e.g., unobserved reputation or quality measures) is particularly problematic in social network settings where the researcher commonly works with cross sectional data on a sparse network that exhibits considerable degree heterogeneity (see, e.g., Chandrasekhar, ; Graham, ). Because the sparsity of the network makes standard fixed effects approaches problematic within these settings, alternative methodologies have been developed for dealing with the identification issue introduced by unobserved degree heterogeneity (see, e.g., Graham, ; Kim, ).…”

Does ethnic diversity decrease economic interactions? Evidence from exchange networks in rural Gambia

“…In fact, we show that if kinship ties are not considered in the analysis, the estimated coefficients are biased towards a negative effect of ethnicity in economic exchanges . Finally, given that we have an almost complete census of links in each village, our estimates are not biased as a result of the selection of a sub‐sample of village respondents (Chandrasekhar & Lewis, ).…”

Politician Family Networks and Electoral Outcomes: Evidence from the Philippines

“…12 Fernandez (2012) describes the data in more detail. 13 Importantly, we do not use a sampled network to generate our centrality measures and as such they do not suffer from the problems identified by Chandrasekhar and Lewis (2016).…”