Study On The Influence Of Analyst Recommendation Changes On Listed Company's Stock Price

Advocating the idea of value investment, Security analysts provide valuable research reports to investors, guiding investors to make investment decisions. Because of the weak-form market efficiency, analysts play a critical role in alleviating the information asymmetry of our security market and in promoting the efficiency of our security market. The security analyst, by collecting and analyzing information, make forecasts and recommendations for the listed corporations, and release their research reports to investors, they might have an impact on the stocks which they issue on. By using average stock-price reactions, the existing domestic literature is not appropriate to measure the influence of analyst recommendation changes. Do recommendation changes from analysts have impact on the stock's price? How often do analysts affect the stock price by releasing research reports? What factors could make them more possibly to do so? We try to answer these questions by empirical analysis in this Paper.The first section of this paper provides an introduction of our research.To examine the questions interested in, firstly, we review the related literature about the influence of analyst recommendation in the second section, the existing literature focused on the average effect of analyst recommendation. They tried to measure the contribution of analyst recommendation through an event study or constructing a portfolio according to analyst' recommendation. They found that analyst do have influence on the firm's stock price. A part of these literatures even go further to examine what factors could affect the analyst's ability to significantly impact the stock's price, they argue that some characteristics such as analyst's reputation, broker's reputation and asset size, the company's size and BM ratio and so on do have effects on the ability of analyst to affect the company's stock price. We will conduct our research on some of their ideas.In the third section, we describe several methods needed later and prepare for the samples through which we are going to have an insight into the influence of analyst recommendation changes. We decide to focus our research on recommendation changes based on prior study results from other scholars. They find that recommendation changes, especially those non-zero ones, are more informative and thus will have a bigger impact on the stock price.In the fourth section, we provide some descriptive statistics of our sample. We find analyst recommendation changes are disproportionally distributed in different magnitude. There is an optimistic mood among analysts that they have a tendency to issue more positive recommendation on stocks in our country. The statistics of average cumulative abnormal return (CAR) shows that analysts do have an impact on stock price roughly. We also find "post-recommendation drift" in CAR, and different magnitude of changes has a different stock reaction. We call a recommendation change influential when it affects stock visibly. We adopt two ways to define being influential and show that less than one thirds of the recommendation changes are influential in our new definition. We find that influential recommendation changes could also have strong impact on other stock characteristics such as volatility, turnover and institutional ownership.In the fifth section, we conduct an independent sample t test to examine the difference between influential and non-influential recommendation changes. We show some recommendation side and firm level characteristics of influential changes are dramatically different from non-influential ones.In the sixth section, we run a Probit model to show that recommendation changes are more likely to be influential when they are from star analyst, bigger broker, issued away from consensus, issued on growth, small and low analyst attention level companies. We find that recommendation changes from star brokers are not more influential than those from non-star brokers. We propose a descriptive statistics to interpret this anomalous outcome. The robustness test shows that our main conclusion is robust to both different samples and alternative definition of influential.Finally, in section seven, we make a conclusion and propose three advices to investors, government and analysts according to our main conclusions.While there are still many valuable things about analyst recommendation changes can be further studied, the main innovation of this paper might be that based on recommendation changes and firm level, we introduce two methods to classify our sample, and take a T test and Probit model to examine what factors make a recommendation change more possible to be influential. This approach can provide a new perspective on the study of the influence of analyst recommendation in our country.