【作者】 胡娜；

【导师】 杨丹；

【作者基本信息】 西南财经大学 ， 财务管理， 2014， 博士

【摘要】 证券分析师作为资本市场上重要的信息中介,通过收集、分析市场、行业和公司层面信息,形成研究报告,发出盈余预测和股票评级,以作为投资者投资决策的参考依据。与西方发达市场相比,我国资本市场具有“新兴加转轨”的特征,这种特殊的制度背景影响到证券分析师行业的发展轨迹。九十年代初以来,我国证券分析师行业经历了一个从无到有、从小到大的发展历程。供职于证券公司和独立咨询机构的分析师,从多角度提供研究报告,降低外部投资者与上市公司之间以及不同投资者之间的信息不对称,其在证券市场上的信息中介的作用日益强化。以“新财富最佳分析师”评选为代表的市场评价机制更是激励了分析师对声誉的追求。另一方面,由于中国特殊的制度背景以及较短的证券市场发展历史,我国分析师在执业过程也面临诸多困境,如不合理的业绩考核体系使得证券分析师过多考虑受雇券商的利益,激励扭曲导致分析师发布有偏的研究报告,从而损害了研究的独立性；监督机制不健全和行业自律机制不完善,以及激励扭曲也可能导致明星分析师为了短期利益而牺牲研究质量；此外,由金字塔控股结构带来的股权复杂性往往造成上市公司信息披露不完善。出于掏空(tunnelling)动机,终极控制人不愿意过多披露信息,从而削弱了分析师从公司管理层获取公司特质信息的能力；上述问题的存在不利于分析师发挥有效信息中介的作用,降低了他们提高资本市场信息传播的效率。本文重点从利益冲突、外部声誉和研究能力三个维度,对我国证券分析师的行为进行研究,并试图探寻背后的机理和原因。理论上讲,资本市场信息效率的发挥有赖于证券分析师独立、客观、公正地研究并传播与目标公司有关的各类信息,如市场信息、行业信息和公司层面信息。现实世界中,卖方分析师很难做到完全独立性。一方面,分析师往往面临来自券商内部激励扭曲和外部市场评价的各种利益纠葛。在利益诱惑和压力胁迫下,分析师往往主动或被动地牺牲独立性而发布偏乐观的报告(Michaely and Womack,1999)。本文的第一部分研究关注券商的股权投资对分析师激励扭曲的影响。重点关注在券商退出股权投资时,旗下分析师是否会发布乐观研究报告推高目标公司的股价,从而攫取更高的投资收益。结果表明：券商的确会把股权投资业务收入的压力传递给旗下分析师,分析师倾向于发布乐观的盈余预测和股票评级。另一方面,处于自身长远职业发展的需要,证券分析师有建立和维护声誉的动机。声誉也被认为是缓和委托代理利益冲突的有效机制(Fama,1980),故本文接下来探讨声誉机制是否起到了约束明星分析师利益冲突的作用。通过考察明星分析师上榜前后预测行为的变化,以新财富最佳分析师评选为例,本文发现明星分析师在被评选为“最佳分析师”后,为了短期利益,存在变现声誉资本的动机,表现为预测质量特征(准确性、一致性和大胆性)三个维度指标的下滑。这说明,声誉未能有效约束明星证券分析师面临的利益冲突。出人意料的是,“声名鹊起”却把分析师置身于更大的利益漩涡中,使明星分析师抵挡不住利益诱惑从而牺牲研究质量的独立性。在前两部分探讨分析师行为受外部客观因素制的基础上,第三部分实证从证券分析师信息挖掘能力入手,以中国上市公司终极控制人的股权结构为切入点,探讨证券分析师对不同终极控股结构公司(金字塔控股结构VS.非金字塔控股结构)信息效率的差异。本文发现金字塔控股结构公司的股价同步性高于非金子塔控股结构的公司,可能的原因是分析师面对复杂的股权结构无能为力,过于复杂的股权机构会加大外部投资者和证券分析师获得公司特质信息的难度,也从另一方面说明证券分析师对公司特质信息挖掘能力有限,从而使得分析师更加倚重市场和行业信息,从而推高该类公司的股价波动同步性。本文共分为七章,各章的主要内容安排如下：第一章为导论,主要介绍研究背景、研究目的、研究内容和论文结构。第二章介绍了美国和中国证券分析师行业的相关制度背景。首先回顾以美国为代表的发达资本市场证券咨询行业的发展历程。在此基础上对中国卖方证券分析师发展的历史和现状进行简述,并比较了两国对证券分析师行业在监管层面的差异。最后对以“新财富最佳分析师“评选活动为代表的买方评价体系进行了重点介绍。第三章为文献回顾,重点回顾了国外学术界对分析师研究报告独立性、分析师预测特征及其经济后果、分析师声誉及其影响、机构投资者对分析师预测偏误的影响、以及证券分析师跟进与市场信息效率等主题的文献。最后,对国内证券分析师的相关研究进展进行了回顾和述评。在充分了解制度背景和以往研究的前提下,论文进入理论和实证研究部分。第四章基于券商股权投资背景,对分析师研究报告的独立性进行拓展研究。考察重点具有股权投资背景的券商在股权投资退出阶段,是否利用旗下证券分析师的“乐观偏见”报告推高目标公司股价,攫取超额投资收益。在第四章的研究基础上,第五章探讨在西方市场上行之有效的声誉机制能否起到约束中国证券分析师利益冲突的作用。在非明星分析师和明星分析师样本配对的基础上,利用倍差法研究明星分析师上榜前后预测行为的变化,研究发现明星分析师上榜后预测准确性和一致性较上榜前显著下滑,这说明明星分析师不存在长期维护职业声誉的动机,我国明星证券分析师的专业能力的可持续性令人质疑。第六章的实证研究从信息挖掘能力的角度探讨证券分析师在我国资本市场上的信息角色。理论上,证券分析师主要向投资者贡献三类信息：市场信息、行业信息和公司特质信息。证券分析师对三类信息的贡献比重取决于三类信息的获取难度。本章以公司股权结构为切入点,研究证券分析师是否有能力克服因股权结构的复杂性所带来的信息获取困难,通过研究金字塔控股结构公司和非金字塔控股结构公司的股价同步性的差异,本文发现证券分析师对公司特质信息挖掘能力有限,证券分析师往往避重就轻,对股权结构复杂的公司更加倚重市场和行业信息,从而推高了金字塔控股结构公司的股价同步性。第七章为研究结论、启示、创新和局限。本章对全文的主要研究结论进行了总结,并在此基础上提出了一系列的政策建议和启示,最后指出了本文的研究创新、贡献与局限性。本文的创新和贡献如下：第一,本文的研究进一步深化了对分析师行为的认识,使我们更加全面地认识投资银行业务对分析师行为的影响,对加强投资者保护具有重要意义。投资银行的业务主要包括承销(1Jnderwriting)、证券经纪交易(Brokerage)、股权投资(PE)。根据防火墙(Chinese Great Wall)原则,证券分析师的荐股与预测行为应该独立于这三项业务。然后,大量研究(Hansen和Sarin,1998; Michaely和Womack,1999; Malmendier和Santhikumar,2009;原红旗和黄倩茹,2007；潘越等,2011)主要从承销关系对分析师行为的角度进行探讨,主要结论是：承销关系对分析师荐股和预测活动会产生重要影响,证券分析师更愿意对其隶属券商承销过的股票给出更为积极的评价。由于数据的披露日益增多,近年来的学术研究也关注到分析师与证券经纪交易之间的关系。如Firth et al.(2013)研究了基金分仓对关联分析师行为的影响,他们发现券商分析师会对券商基金客户所持股票给出较为积极的投资评级。这种乐观性倾向会随着所荐股票在基金投资组合中的权重以及交易佣金的增加而增加。这些研究深化了我们对防火墙问题的认识,但由于数据限制,目前却缺少股权投资业务分析师行为影响的研究。基于中国市场的独特性,利用手工搜集到的数据,用分析师所属券商在目标公司Pre-IPO前是否进行股权投资作为是否存在关联关系的刻画,本文对这一问题进行了深入探讨。研究结果表明,在解锁期以前,由于关联分析师存在信息优势,因此其预测行为的精确性较高。在解锁期之后,关联分析师更倾向于给出乐观型评估,从而便利于其所属券商在退出股权投资之际获得较高的投资收益。借鉴Kothari (2001)等人的说法,这反映出关联关系对分析师活动产生战略性扭曲(strategic distortion)作用。分析师在进行荐股或盈余预测时,会自觉地考虑到利益关系的重要性,从而做出适当调整。第二,本文的研究丰富了人们对明星分析师声誉评价体系(Superstar system)所带来的激励问题(Rosen,1981,1992; MacDonald,1988; Andersson,2002; Krueger,2005;)的认识。作为资本市场运行的基础,声誉起着至关重要的作用,但由于数据的缺乏,已有研究多为理论探讨。鉴于《机构投资者》杂志公布的全美明星分析师(institutional investor all-american star equity analysts)在美国资本市场上具有重要影响力,因此分析师声誉方面的相关研究多以此为背景(Stickel,1992,1995; Desai et al.,2000; Fang和Yasuda,2004,2009; Li,2005; Leone和Wu,2007等)。这些研究深化了我们对分析师声誉问题的认识,但已有研究较少涉及排名所带来的激励问题。比如,由于最佳分析师声誉给相关分析师带来收入上的巨大提升和行业知名度的提高,其荐股和预测行为是否因此而发生扭曲?本文的研究结果表明,明星分析师在上榜前的预测表现要普遍优于非明星分析师,这说明较高的能力使他们获得较高的声誉。但中国明星分析师在上榜以后,其预测质量却呈现明显下滑趋势。这说明中国明星分析师不存在维护职业声誉的动机。换句话说,中国明星分析师的专业技能的可持续性令人生疑。第三,本文的研究加深了对分析师信息传播效率的认识。分析师跟进能否有效地传播公司层面的特质信息是学者们长期关注的问题。事实上,国内已有研究(朱红军等,2007,冯旭南和李心愉,2011)已经关注到该问题,但两者得出的结论并不一致。尽管样本区间和研究方法的差异是造成研究结论相异的原因,但未考虑分析师对信息挖掘能力也可能是导致结论差异的重要因素。我国上市公司复杂股权结构所带来的信息壁垒是考察证券分析师信息挖掘能力的一个良好视角。以金字塔控股结构的公司为例,该类股权结构的公司受终极控制人对内部信息的控制,为外部投资者和证券分析师制造了较高的信息壁垒。因此,以金字塔控股结构为背景,通过深入研究分析师跟进与股价同步性之间的关系,考察证券分析师对公司特质信息挖掘的能力。该部分的研究表明：金字塔控股结构导致分析师难以获取公司层面的特质信息,因此其股价同步性较高。另一方面,该部分的研究深化了我们对股价同步性问题的认识。因为Chan和Hameed (2006)指出,产权保护程度的差异可能是导致不同国家股价联动性差异的重要原因。以中国为背景,从终极控制人的掠夺动机出发,以金子塔控股结构为切入点,该部分探讨金字塔结构和非金字塔结构对分析师传播公司特质信息能力的影响,因此丰富了国外学界(Wurgler,2000; Piotroski和Roulstone,2004; Chan和Hameed,2006)对股价同步性问题的认识。更多还原

【Abstract】 As important information intermediaries in capital market, securities analysts collect and analyze market, industry and firm-specific information related to followed listed firms. They inform investors with their judgment on firm prospect in form of earnings forecasts and stock recommendations. Compared with western developed markets, an emerging capital market like China has shaped the trajectory of securities analyst professionalism. The institutional arrangements and brief development history have put Chinese securities analysts into a series of predicaments. For instance, a variety of conflicts of interests compel analysts to issue upward biased research reports and thus compromise their research independence; It is difficult for star analysts to maintain superior forcasting level, and their professional reputation is questionable due to lack of regulation and supervison; Complex corporate ownership structure, such as pyramidal ownership structure, leads to opaque information disclosure, which challenges analysts’ ability to access and acquire firm-specific information.My dissertation tends to study securities analysts’ behavior from three dimensions like research independence, conflicts of interests and professional competence. Theoretically, the information efficiency of capital market relies on the dissemination of objective, timely and unbiased information by securities analysts. In reality, the independence of sell-side analysts hardly sustains. Analysts submit themselves to a variety of conflicts of interests and issue optimistic research reports (Michaely and Womack,1999). The first issue we are interested in is whether securities broker would pressure analyst employees to issue optimistic reports in order to lift stock price when withdrawing equity investment. Empirical results show that analysts fail to resist the pressure from brokers and issue optimistic earnings forecasts and stock recommendations. On the other hand, concerning long-term career development, analysts have incentives to establish and maintain reputation. Reputation has long been viewed as an effective mechanism to alleviate conflicts of interest between principle and agent (Fama,1980). Therefore, the second issue we examined is to discuss whether reputation curbs analysts’ optimism by observing forecasting behavior prior to and after getting listed as stars. Consistent with Fang and Yasuda (2009), we find that forecasting quality indicators deteriorate after being voted as stars, suggesting star analysts tend to liquidate their reputation capital. It’s difficult for analysts to maintain professional reputation over long run. The malfunctioning of reputation drives us to further explore underlying mechanisms behind the scene. We wonder if unique features of Chinese institutional setting affect analysts’ information efficiency. Specifically, we focus on stock synchronicity of listed firms with different ownership structure (pyramidial structure vs. non-pyramidial structure). Empirical results show that the stock synchronicity of firms with pyramidal ownership is higher than that of firms with non-pyramidal ownership. The underlying reason is that the complexity of ownership structure provides ultimate controllers with opportunities of manipulating inside information, raising the threshold for outside investors and analysts to access firm-specific information. Consequently, analysts have no option but to rely more on market and industry information, uplifting stock synchronicity of firms with pyramidal ownership. On the other hand, professional competence of analysts in China needs to be improved. They seem to be unable to collect and disseminate firm-specific information on firms with more complex structure ownership.This dissertation consists of seven chapters, each chapter proceeds as follows:Chapter one is introduction, describing research background, objectives, content and structure.Chapter two summarizes institutional background of securities analyst professionalism in United States and China, respectively. Based on introductions on the history and current status of sell-side securities analysts, the discrepancies on securities analyst professionalism between United States and China are discussed. Finally, the voting event of "New Fortune Best Analyst", a market-oriented performance evaluation system, is introduced briefly.Chapter three focuses on literature review. Overseas literature on financial analysts is reviewed along the dimensions of forecasting characteristics and economic consequences, analysts’ reputation and its influence, the influence of institutional investors on forecasting bias as well as analyst following and information efficiency. Ultimately, related domestic research on financial analysts is reviewed and potential research fields are identified.Chapter four extends the research on analysts’ conflicts of interests. Taking the perspective of securities brokerage firm with equity investment, chapter four examines whether if brokers pressure analysts to lift stock price of target firms by issuing optimistic research reports when it comes to withdrawing equity investment. The empirical results support the hypotheses.Chapter five explores whether reputation functions as well in Chinese stock market as in western stock market to curb analysts’interests of conflicts. Matching stars with non-stars, we use difference-in-difference (DID) to study analysts’ forecasting behavior before and after getting listed. We find that forecasting accuracy and consistency deteriorate after analysts are voted as stars, suggesting that it’s difficult for star analysts to sustain superior professional competence.Chapter six evaluates securities analysts’competence from the perspective of firm-specific information mining. Presumably, analysts contribute three categories of information to investors:market, industry and firm-specific information. The contributing proportion depends on the difficulty of information acquisition for each category. In this chapter, we attempts to examine whether different ownership structures affect analysts’initiatives to dig into firm-specific information. Specifically, we find that firms with pyramidal ownership exhibit higher stock synchronicity than counterparts with non-pyramidal ownership.Chapter seven presents research conclusions, policy implications, innovations and limitations.The research innovations and contributions are as follows:First, different from prior studies on how investment banks support equity investment business line at entry stage(Michaely and Womack,1999; Fang,2012), the first study, taking the perspective of withdrawing equity investment, examines how brokers push affiliated analysts to issue optimistic reports and gain extra investment proceeds. In this study, we expand categories of equity investment to increase the variety and size of sample, making research conclusions more convincing. Second, the current study enriches the line of literature on analysts’ reputation mechanism.In particular, the second study extends Fang and Yasuda (2009) and finds consistent evidence with "analysts’ reputation liquidation hypotheses", which deepens our understanding towards analysts’ reputation. The results suggest that the voting of New Fortune Best Analysts does have some merits. The star list at least partially reflects analysts’ research abilities as reflected by forecasting accuracy and consistency. The research level of non-stars improves over time, suggesting the motivating effect of this voting system. On the other hand, the second study also opens up a new question for future research, that is, why star analysts tend to liquidate reputation at the price of compromising research quality.Third, the current study makes contributions to the line of research on analysts’ information efficiency. Although prior studies have investigated analyst information dissemination efficiency (Fan et al.2002; Feng and Li,2013), few researchers examine this issue from the perspective of analysts’ ability of information mining. The current study attempts to fill this void by introducing ultimate ownership structure. We find that the complexity of corporate ownership impedes analysts from getting access to firm-specific information and uplifts stock synchronicity. In other words, analysts’ information mining ability is limited due to the complexity of corporate ownership structure.更多还原