| Summary: | 碩士 === 國立交通大學 === 財務金融研究所 === 104 === This study mainly discusses the relationship between investor sentiment and analysts price estimate. Not like previous research targeting stock recommendations or estimate of earnings per share, we’ll target analyst price estimate instead. Our target-price-related dependent variables are Price Forecast Update Index, Price Forecast Change Index, Signed Error and Absolute Error. As for sentiment proxies, we select sentiment index derived from Baker Wurgler (2006) (BW), Consumer Confidence Index of Michigan University (MCCI), and Conference Board’s Consumer Confident Index (CBCCI). Our motive is based on Bradshaw, Brown and Huang (2013) using price estimate as our main variable and we’ll follow the research process of Kaplanski and Levy (2014), trying to find the causality of sentiment and analysts price estimate.
Our main findings are presented as follows. Firstly, in our time-series analysis we find that analyst price estimate will become more optimistic as the sentiments are positively increasing. It comes with that signed error and absolute error will also increase after positive change of sentiment. Secondly, by conducting Granger Causality Test, we find statistically significant result that the change of CBCCI can predict the change of estimate error, not vice versa. Furthermore, we can tell from our panel data that more profitable firms are more strongly affected by the change of sentiment, in other words, estimate errors will significantly increase when sentiment is positively moving in those better-doing firms. But we cannot find the same results in other firm characteristics, such as the number of price estimate of a firm, size, volatility, or stock price momentum. Last but not least, even to the very little extent, we find that analysts are not immune to the change of some macroeconomic variables.
|
|---|
