证券分析师是资本市场重要的信息中介，是降低投资者和上市公司内部人之间信息不对称的重要信息媒介。他们发布的研究报告，包括盈利预测、荐股以及报告的分析文本，一直是投资者开展投资决策的重要依据。当市场存在较大不确定时，投资者有更迫切的信息需求，而作为信息中介的证券分析师如何回应这一需求，是一个重要的研究课题。另外，证券分析师倾向于向市场提供偏于乐观的投资建议，但是这种偏差是源于证券分析师自身的认知缺陷还是机会主义的自利动机，尚有很多探讨空间。现有文献对证券分析师如何应对不确定性、以及证券分析师偏误来源的研究很大程度受到内生性的影响，本文利用新冠疫情这一外生事件为研究背景，以实证研究的方法探讨这两个问题。

而新型冠状病毒肺炎疫情（以下简称“新冠肺炎疫情”）是一种重大突发公共卫生事件。新冠肺炎疫情作为背景，具有高度外生性，包括证券分析师在内的所有市场参与者都无法准确预测这一事件的发生、开始时间、事件进程以及对经济活动的影响，证券分析师也无法据此决定自己跟踪哪些行业、哪个公司，即新冠这一事件对分析师预测活动的影响不会受到证券分析师本人事先偏好的影响。因此，新冠肺炎疫情带来的巨大不确定性为我们研究证券分析师的盈利预测行为提供了理想的研究背景。

本文的研究目标是以新冠肺炎疫情为研究背景，探讨在面临巨大的不确定性情况下，证券分析师盈利预测行为的特征。具体而言，当整个宏观环境受到外界冲击、面临巨大的不确定性时，证券分析师的盈利预测准确性和及时性会产生怎样的变化？证券分析师的盈利预测偏差又将受到怎样的影响？新冠肺炎疫情对证券分析师盈利预测行为的影响是否会受到自身能力差异的影响？

本文首先通过描述性分析了解数据的基本特征和变量之间的相关关系，发现新冠肺炎疫情前后证券分析师的盈利预测行为有显著差别，且变量之间具有高度相关性。再通过基准回归和交互项回归，对假设进行验证，分别得出以下结论：面对新冠肺炎疫情带来的巨大不确定性，（1）证券分析师盈利预测的准确性显著提升；（2）证券分析师盈利预测的及时性显著下降；（3）证券分析师的盈利预测会更加悲观；（4）证券分析师工作年限会对盈利预测准确性造成显著正面影响，降低乐观偏差程度；（5）证券公司净资产规模、发布研究报告数量会对证券分析师的盈利预测准确性造成显著正面影响，但不会影响乐观偏差；（6）证券分析师个人能力和证券公司质量均会对及时性造成负面影响。

本文基于以上初步研究结论，从自然状态下不同程度的不确定性（特殊行业、具有我国特色的上市公司板块）和人工干预后的不确定性（防疫政策）入手，开展进一步研究。以新冠肺炎疫情引起巨大需求的医药制造行业作为研究变量发现，证券分析师对医药制造行业上市公司盈利预测更加频繁、反应更加乐观。以注册制试点的科创板作为研究变量，发现证券分析师盈利预测的准确性更低、乐观偏差更大。以首个执行“封城”政策的湖北省为例，通过DID模型、PSM匹配和稳健性检验，发现“封城”政策会对新冠肺炎疫情背景下证券分析师的行为特征有加剧作用，即进一步提高准确性降低及时性，悲观反应会更激烈。

本文的研究特点在于以高度外生性的新冠肺炎疫情作为背景，研究证券分析师的盈利预测行为，尤其是乐观偏差。本文以新冠肺炎疫情为背景，避免了内生性问题，是对现有研究证券分析师行为特征文献的有力补充。本文发现，面临巨大的不确定性下投资者对信息准确性和及时性的更高要求时，证券分析师选择耐心等待并尽可能为投资者提供更准确的盈利预测，说明证券分析师更关注其信息的准确而愿意一定程度上牺牲及时性。此外，本文发现在面临巨大的不确定环境下，证券分析师并未如现有文献发现的，发布过度乐观的投资建议，而是发布了更偏悲观的盈利预测，说明证券分析师先验的主观偏误是其过度乐观预测行为的重要原因之一，而非全部由分析师的自利投机动机造成。

Financial analysts are important market participants and information intermediary on the capital market. Their research reports, including quantitative outputs such as EPS forecasts and stock recommendations, and qualitative outputs such as their textual analyses on the industry competition and firm strategy, are important information sources for investors’ investment decisions. This study examines how analysts react to the high uncertainty and whether they issue biased forecasts in highly uncertain conditions using the exogenous setting of COVID-19 pandemic.

  And the new coronavirus pneumonia epidemic (hereinafter referred to as“COVID-19 epidemic”) is a major public health emergency. The COVID-19 epidemic as a background is highly exogenous, all market participants, including financial analysts, can not accurately predict the event's occurrence, timing, course, and impact on economic activity, nor can financial analysts decide which industries and which companies they follow. The impact of covidOVI-19 on analysts' forecasts will not be influenced by their own predilections. Therefore, the great uncertainty brought by COVID-19 provides an ideal research background for our research on the EPS forecasts behavior of financial analysts.

The objective of this study is to explore the characteristics of EPS forecasts behavior of financial analysts in the face of great uncertainty, with the COVID-19 epidemic as the research background. Specifically, when the overall macro environment by external shocks, faced with great uncertainty, how will the accuracy and timeliness of financial analysts' EPS forecasts change? How will the EPS forecast deviation of financial analysts be affected? Will the impact of COVID-19 epidemic on analysts' EPS forecast be affected by their own ability differences? 

Firstly, this paper analyzes the basic characteristics of the data and the correlation between the variables by descriptive analysis, and finds that there is a significant difference in the EPS forecasts behavior of the financial analysts before and after the COVID-19 epidemic, and the variables are highly correlated. Through the benchmark regression and the interactive term regression, the hypothesis was tested, and the following conclusions were drawn: faced the great uncertainty brought by COVID-19, (1) the accuracy of financial analysts’ EPS forecasts has improved significantly; (2) the timeliness of financial analysts’ EPS forecasts has decreased significantly; (3) financial analysts’ EPS forecasts have become more pessimistic; (4) the working years of financial analysts will have a significant positive impact on the accuracy of EPS forecasts, reducing the degree of optimism bias; (5) the net asset size of securities companies and the number of research reports will have a significant positive impact on the accuracy of EPS forecasts, but will not affect the optimism bias; (6) the personal ability of financial analysts and the quality of securities companies will have a negative impact on timeliness..

Based on the above preliminary research conclusions, this paper starts further research from the natural state of different degrees of uncertainty( the special industries, listed companies with our characteristics) and artificial intervention after the uncertainty( epidemic prevention policy). Taking the pharmaceutical manufacturing industry, which is in great demand due to COVID-19, as the research variable, the empirical test shows that the financial analysts’ EPS forecasts are more frequently and optimistic. We take the STAR Market that make a pilot study of registration system as the research variable, and find that the accuracy of EPS forecasts is lower and the optimistic deviation is bigger. This paper takes Hubei province as an example to further study the impact of the “lockdown the city” policy on the EPS forecasts behavior of financial analysts through the DID model, PSM matching and robustness test. It is found that the policy would lead to improving accuracy and reducing timeliness, and more pessimistic response.

The characteristic of this paper is to study the EPS forecasts behavior of financial analysts, especially the optimism bias, against the background of the highly exogenous COVID-19 epidemic. Based on the COVID-19 epidemic, this paper avoids endogenous problems and is a powerful supplement to the existing literature on behavior characteristics of financial analysts. This paper finds that when faced with the higher demand of accuracy and timeliness of information under great uncertainty, analysts choose to be patient and provide investors with more accurate EPS forecasts as far as possible, it shows that analysts are more concerned about the accuracy of their information and are willing to sacrifice timeliness to some extent. In addition, the paper found that in an environment of great uncertainty, analysts did not issue overly optimistic investment recommendations, as the literature has found, but issued more pessimistic EPS forecasts, it shows that the subjective bias of analysts is one of the important reasons for their over-optimistic forecasting behavior, and not all caused by analysts’ self-interest speculation motive.