Research On The Securities Investment Ability Of Analysts And Fund Managers Based On Quantile Regressions

Although China's securities market started later than the developed capitalist countries,it has achieved world-renowned achievements after decades of development.The value concept and investment philosophy in Chinese securities market has gradually been recognized and accepted by investors and plays an important role in the capital market.The ability of analysts and fund managers' investment choice is exist or not has drawn more and more attentions from academics and practitioners.In recent years,domestic and foreign scholars have widely discussed the investment choice ability of analysts and fund managers and got fruitful research results.However,the empirical results and findings mainly depend on mean regressions that are unable to comprehensively describe the heterogeneity of securities investment choice ability.The quantile regression method proposed by Koenker et al.overcome the shortage of mean regression and provides a basic tool solve such problems.In this regarding,we apply quantile regressions to study the investment selection ability of securities investors and to explore the heterogeneity and variation in different levels.In this dissertation,we conduct empirical analysis from the following two aspects:(1)We study the influence of non-financial factors on the forecasting ability of securities investment analysts.Based on non-financial information,we apply quantile regressions to examine the influence of corporate social responsibility information disclosure on analyst prediction ability.In addition,we throughly investigate the heterogeneity influence of enterprise information disclosure quality on analyst forecast dispersion,analyst forecast accuracy and analyst following.The empirical analysis is conducted on Chinese A-share companies and the results are promising.The empirical results show that the information disclosure of corporate social responsibility can improve analyst's forecasting ability in an effective way,the quality of the information companies disclosed can attract analysts to follow and improve the analyst forecast accuracy and reduce the analyst forecast dispersion.Interestingly,we find that the degree and way of the effect are relate to the market environment(quantile-dependent).(2)We evaluate the ability of securities investment fund manager on timing and stock selection.Under the framework of quantile regression,we extend the three-factor T-M model with generalized additive structure and propose a generalized additive quantile regression model to evaluate the ability of fund managers on timing and stock selection in an more effective way.We conduct empirical analysis on Chinese stock market and draw the following conclusions.First,the generalized additive quantile regression model is able to accurately evaluate the ability of fund managers on stock selection and timing.The abilities are heterogeneous across different quantiles.Second,the fund managers in Chinese stock market are generally short of security-selecting ability,but have a certain timing ability.Third,the ability of Chinese fund managers will be better at a higher quantile,which means that a better investment environment can improve the ability of fund managers.To sum up,we develop models of securities investment choice ability under the framework of quantile regression,which enriches the traditional econometrics models theoretically.The novel models help to explore the heterogeneous effects that can not be found with the traditional models.The empirical results and research conclusions of this dissertation can provide decision support for investor,fund company,and supervision department to prevent financial risk.