顶级券商的明星分析师荐股评级更有价值么?——基于券商声誉、分析师声誉的实证研究

利用2003—2009年《新财富》杂志评选作为分析师声誉和券商声誉代理变量,结合超过54000条分析师推荐评级样本,从投资者角度检验了券商声誉、分析师声誉与分析师荐股评级价值之间的短期关系和长期关系,研究结果发现:在评级公告日,"买入/增持","中性","减持/卖出"评级分组的超常收益率与分析师声誉和券商声誉之间存在不同的影响规律。通过长期关系检验还发现牛市阶段,非顶级券商明星分析师"买入/增持"评级的投资价值高于顶级券商明星分析师,但在熊市阶段,非顶级券商非明星分析师"买入/增持"评级投资价值最高,顶级券商明星分析师投资价值最低。顶级券商和非顶级券商的明星分析师"中性"评级投资价值低,投资者容易错失投资机会。顶级券商明星分析师"减持/卖出"评级投资价值最高;非顶级券商明星分析师"减持/卖出"评级投资价值最低。最后提出了维护分析师证券研究报告独立性的四条政策建议。

Do Star Analysts’ Recommendations from Top Brokers have more Investment Value? An Empirical Study Based on Reputation

Using the New Fortune Magazine′s star analysts ranking as the agent of reputation,combined with analysts’ stock rating samples exceeding 54000,this paper tests short-term and long-term relationship between analyst reputation and the investment value of the recommendation.The results show that the investment value from different groups of recommendation samples take on different correlation with brokers′ reputation and analysts′ reputation. In the first part,the short-term model examines the relationship between analyst recommendation performance and reputation.Market-adjusted return of recommended stock on the announcement day is adopted as the agent of analyst′s recommendation performance to regress on specific explanatory and control variables.If more than one analyst recommendation on a particular stock,then this stock will appear multiple times in the regression,once for each recommendation.Prior research(Barber et al.,2001;Stickel,1995;Womack,1996) has shown that small firms exhibit a greater absolute response to recommendations than do large firms.Since the stock size and broker size samples have a highly positively skewed distribution,the natural logarithm of the market value of recommended stock is used as the first explanatory variable.Barber(2001) found that stock price have bigger reaction to the analyst′s recommendations from bigger broker.The natural logarithm of broker size is used as the second explanatory variable. To test the impact of analyst′s reputation on abnormal return of recommended stock on the announcement day,the analyst′s professional reputation is decomposed into two parts.The first part is the accumulative times being a New Fortune Magazine′s star analyst.The more accumulative times being a star,the higher the analyst′s reputation.The second part is the accumulative ranks in his following industry in recent years.The higher accumulative ranks in one industry,the higher the analyst′s reputation.Similarly,to test the impact of broker′s reputation on abnormal return of recommended stock on the announcement day,the broker′ professional reputation is decomposed into two parts too.The first part is the accumulative times being a New Fortune Magazine′s most influential broker.The more accumulative times being the most influential broker,the higher the broker′s reputation.The second part is the accumulative ranks of broker′s market influence in recent years.The higher accumulative ranks in market influence,the higher the broker′s reputation. The regression results of "strong buy/buy" rating samples show that the absolute value of′ market-adjusted return have positive correlation with the accumulative times being a star analyst but have negative correlation with the star analyst accumulative ranks.The absolute value of "Strong buy/buy" rating samples′ market-adjusted return have positive correlation with the accumulative times being a star broker,but have negative correlation with the accumulative ranks being a star broker.Whether accumulative times or accumulative ranks,the effects of broker reputation on the abnormal return of "strong buy/buy" rating samples are lower than analyst reputation.The regression results of "neutral" rating samples show that the absolute value of′ market-adjusted return have positive correlation with the accumulative times being a star analyst but have negative correlation with the star analyst accumulative ranks.The absolute value of the market-adjusted return of "neutral" rating samples have negative correlation with the accumulative times being a star broker,but have positive correlation with the star broker accumulative ranks.Whether accumulative times or accumulative ranks,the effects of broker reputation on the abnormal return of "neutral" rating samples are similar with analyst reputation.The regression results of "strong sell/sell" rating samples show that the absolute value of market-adjusted return have negative correlation with the accumulative times being a star and accumulative ranks.The absolute value of the market-adjusted return of "strong sell/sell" rating samples have negative correlation with the accumulative times being a star broker but have positive correlation with the star broker accumulative ranks.Whether accumulative times or accumulative ranks,the effects of broker reputation on the abnormal return of "strong sell/sell" rating samples are higher than analyst reputation. In the second part,in order to determine whether the analyst′s reputation and broker′s reputation are useful in predicting the performance of recommendations,the paper turns to an examination of whether longer-term recommendation returns differ across analyst reputation,broker reputation and bull/bear market.We begin our analysis by calculating,for each group,returns to each of our 27 separate recommendation subsamples.If more than one analyst recommendation on a particular stock,then that stock will appear multiple times in the portfolio,once for each recommendation.Assuming an equal dollar investment in each recommendation,the dynamic portfolio is constructed.The long term model discusses the relationship between analyst recommendation performance and broker/analyst reputation.Three measures of risk-adjusted performance are calculated for each of our portfolios.The first is the mean daily market-adjusted return,found by subtracting the daily return on the value-weighted market index from the daily return of each of our portfolios.The second is the intercept from the three-factor model developed by Fama-French(1993).The third is the intercept from the four-factor model developed by Carhart(1997). It is found that star analysts’ "Strong buy/buy" ratings from non-top broker have more investment value than star analysts from top brokers in bull market.But non-star analysts′ "Strong buy/buy" ratings from non-top brokers have the largest investment value in bear market.Star analysts’ "Strong buy/buy" ratings from top brokers have least investment value in bear market.It is found that star analysts’ "neutral" ratings from top or non-top brokers have low investment value,investors may miss the investment chances.It is found that star analysts’ "Sell/Strong Sell" ratings from top brokers have the largest investment value.Star analysts’ "Sell/Strong Sell" ratings from non-top brokers have the least investment value. In summary,analyst reputation and broker reputation both have reference value for investors to judge the investment value of ratings.But star analysts from top brokers do not always have superior stock-picking ability than other normal analysts.The conclusion of this paper provides direct evidence of defectiveness of New Fortune Magazine Star analyst evaluation mechanism.With the lack of efficient supervision and analyst reputation formation and punishment mechanism,the sell-side analysts market will eventually become a lemon market.The opportunistic inferior analysts will expel those high-capability and high-ethics analysts. Four policy suggestions are put forward to improve the independence of analyst research and analyst supervision at the end of the paper.First,election and standards of star analysts by media should be supervised by China Securities Regulatory Commission and Securities Industry Association.Second,internal control system of broker′s research department should be improved.Third,quality control mechanism of analysts′ research reports should be enhanced.Forth,new technology should be adopted to improve the analyst supervision efficiency.