Attention Networks: A Two-Mode Network View on Valuation

Abstract

When multiple actors allocate their attention across multiple issues, they create an attention network. We leverage the multiple ties that comprise such an attention network to argue that how competitors interpret a given market situation depends not only on information about that situation but also on the portfolio of other situations to which they pay attention. Specifically, we hypothesize that the more two competitors assess a given situation vis-a-vis a more similar portfolio of other situations, the more they assess that issue similarly. It is so, we argue, because i) competitors who pay attention to the same problems are more likely to make similar cognitive associations and consequently come to more similar solutions on a given problem, and ii) competitors are more likely to adapt their interpretations on the basis of the assessments made by the competitors with whom their attention patterns intersect more frequently. We exploit the two-mode (agents-issues) dynamic structure of these observational networks to study the social and temporal aspects of financial analysts’ valuations in markets from 1993-2011. Our findings show that two analysts form more similar earnings estimates about a given firm if they come to that assessment while paying attention to a more similar portfolio of other firms. Our analysis further demonstrates that analysts’ estimates of the focal firm are influenced by views that cycle through closed triads.