文澜学术系列讲座 第八十九期 对外经济贸易大学张书宇老师:“Causal language intensity in performance commentary and financial analyst behaviour”
日期:2017-09-07             信息来源:文澜学院

主题 / TopicCausal language intensity in performance commentary and financial analyst behavior

时间 / Time98号(周五)|September 8 (Friday), 1400- 1520

地点 / Venue文澜401教室WENLAN

  

主讲/ Speaker


张书宇博士,对外经济贸易大学助理教授,博士毕业于比利时安特卫普大学,获得应用经济学博士学位;硕士毕业于荷兰蒂尔堡大学,获得金融学硕士学位。在国内外知名期刊,如Accounting and Business ResearchEuropean Management Journal,《经济科学》等发表多篇文章。

  

研究领域/ Research Interests 


会计与金融,金融学,管理学

  

摘要/ Abstract


本文使用人工智能算法来衡量上市公司年报中管理层讨论部分的因果性信息与分析师行为的相关性。我们发现,(1)公司能够通过提供更多的因果性信息来吸引更多的分析师跟随。并且分析师也能从这些因果性信息中获益,表现为分析师对于净利润的预测更准确,分析师之间对于净利润的预测分歧更小。这些结果说明,因果性信息提供了额外的信息,降低了分析师的信息成本,并且总体上降低了信息环境的不确定性。此外,我们还发现,(1)以上主要发现在针对未来的因果性信息中相关性更为显著。(2)如果跟随公司的分析师下降,公司会在次年的年报中提供更多的因果性信息。(3)分析师如果已经覆盖了很多公司,那么这些分析师将倾向于新覆盖提供更多因果性信息的公司。


We use computer-intensive techniques to measure causal reasoning on earnings-related financial outcomes of a large sample of MD&A sections of US firms and examine the intensity of causal language in that context against extent of analyst following and against properties of analysts’ earnings forecasts. We find a positive and significant association between a firm’s causal reasoning intensity and analyst following and analyst earnings forecast accuracy respectively. Correspondingly, analysts’ earnings forecast dispersion is negatively and significantly associated with causal reasoning intensity. These results suggest that causal reasoning intensity provides incremental information about the relation between financial performance outcomes and its causes, thereby reducing financial analysts’ information processing and interpreting costs and lowering overall analyst information uncertainty. These associations are stronger for forward-looking causal reasoning intensity. Additionally, we find that decreases in analyst following are followed by more causal reasoning on performance disclosure. We also find analysts who already cover many firms intend to follow a firm with more causal reasoning. Overall, our various evidences of causal reasoning intensity and properties of analyst behavior is consistent with the proposition that causal reasoning is a generic narrative disclosure quality characteristic, able to provide incremental information to analysts and possibly guide analysts' behaviour.