Files
Abstract
Given the rich information content of financial news and its usefulness for decision-making in analysts in scenarios like earnings estimates, previously found determinants of analyst forecast revisions are reconstructed using the high dimensional word embedding space and financial news data from the Dow Jones Newswires. By building word analogy in the embedding space, I find evidence of analyst optimism associated with downward forecast revisions. Financial feature dimensions such as positivity and certainty are quantified using sets of antonym word vectors, and this enables us to gauge the relational impact of any interested factors on analyst forecast revisions via the determinant dimensions. Furthermore, strongly positive and significant impact of firm-specific sentiments are found via regression methods. These evidence together suggest that financial news provide incremental informativeness in explaining individual-level analyst forecast revisions, and corroborates the critical roles of analyst optimism and firm-specific sentiments in analyst forecast revision decisions.