Logistic Regression Model for Cognitive Biases in Financial Analysts' Recommendations

Basing on event study method, this paper explores whether analysts are prone to cognitive bias or not using a sample of 1857 recommendations for listed companies during 2011 similar to 2013 made by financial analysts from Shenyin & Wanguo Securities Research in China. Logistic regression results show that financial analysts are likely to award optimistic recommendations to stocks with better prior price performance and lower book-to-market. Moreover, good optimistic recommendations are characterized by better prior price performance, lower book-to-market and smaller firms size. The finding confirms the existence of cognitive biases in analysts' recommendation behaviour.