Is herding spurious or intentional? Evidence from analyst recommendation revisions and sentiment

Abstract

Using institutional holdings data for 1993–2015, we investigate whether institutional herding is spurious or intentional by analysing the impact on herding of analyst recommendation revisions and sentiment. In addition to examining their effect separately, we consider the impact of their interaction. Utilising the Sias method, we develop competing hypotheses concerning spurious and intentional herding, an issue of direct interest to fund investors, given the principal-agent relationship inherent in fund management. Results strongly suggest herding is spurious. Analysis of the relationship between herding and subsequent returns, and findings from robustness tests add further support.