我国A股经历了30年的快速发展已经跻身全球第三大股市，但仍未到达成熟阶段，个人投资者占比过大导致投资者结构失衡从而使得“追涨杀跌”、“羊群效应”等市场非理性现象层出不穷。为解释金融市场上的异象，理解证券的定价逻辑，本文从行为金融学出发，对前人研究进行分析，并改进投资者情绪指数：选取融资余额变化率、换手率、新增开户数、市盈率、投资者信心指数5个变量作为源指标，在考虑源指标对综合指数的“提前”、“滞后”效应后，选取PMI、CPI、PPI、IVA作为宏观经济控制变量剔除变量中的宏观经济影响因素。而后使用主成分分析方法重新构建投资者情绪综合指数，并将其与使用原构建方法构建的指标进行相关性分析。在此基础上，将新构建的投资者情绪指数与股市综合指数以及不同行业指数对数收益率在一个更长的样本区间内使用VAR模型，Granger因果检验，脉冲响应分析，方差检验方法研究投资者情绪与股市收益之间变动关系。结果发现，新构建的投资者情绪指数与使用原方法构建的指数具有较弱的正相关关系，说明新构建的指数效果较好。同时发现，从总体角度看：股市综指对投资者情绪的冲击存在时滞，而后在较长时间内对投资者情绪产生正向冲击，在达到峰值后逐渐降低并趋于0。投资者情绪会对股市综指对数收益迅速产生正向冲击，并在第2期转为负向冲击，达到最小值后冲击效果出现正反交替。从行业角度看，投资者情绪与股市收益的关系变现出行业间的差异性：（1）模型滞后阶数不同，交运仓储、住宿餐饮、信息技术、金融业、房地产、科研技术等10个行业滞后阶数较短；（2）Granger因果关系不同，在5%的显著水平下，只有水电煤气、住宿餐饮、租赁商务、卫生社会4个行业不存在与投资者情绪的双向Granger因果关系；（3）将行业指数划分为公用、周期、TMT和大金融板块，发现公用、周期板块内的行业指数受到投资者情绪的影响较大，TMT板块内的行业指数受到投资者情绪的影响较小。最后提出了加快推进资本市场发展，丰富金融市场可用套利工具，以及加强投资者教育和信息披露以减少投资者非理性行为的建议。
After 30 years of rapid development, China's A-share market has become the third largest stock market in the world, but it has not been mature. The large proportion of individual investors leads to the imbalance of investor structure, which leads to the endless market irrational phenomena such as "chasing up and killing down", "herding effect". In order to explain the anomalies in the financial market and understand the logic of securities’ pricing, this paper analyzes the previous studies, improves the investor sentiment index and selects five variables as the source indexes, including the change rate of financing balance, the turnover rate, the number of newly opened accounts, the PE ratio and the investor confidence index. After considering the "advance" and "lag" effects of the source index on the composite index and using PMI, PPI, CPI and IAV as macro-economic control variables, the principal component analysis method is used to build investor sentiment. We also analyses its correlation with the indicator constructed by the original construction method. On this basis, we use VAR model, Granger causality test, impulse response analysis and variance test to analyze the relationship between investor sentiment and stock market return.The results show that there is a weak positive correlation between the new investor sentiment index and the original index, which shows that the new index is effective. At the same time, from the overall point of view, there is a two-way Granger causality between investor sentiment and the stock market composite index. There is a time lag in the impact of the stock market composite index on investor sentiment, and then the stock index has a positive impact on investor sentiment for a long time, which will gradually decrease and approach zero after reaching the peak. On the other hand, investor sentiment will quickly produce a positive impact on the stock index, which will turn into a negative impact in the second period. After reaching the minimum value, the impact effect tends to zero and appears positive and negative alternation. From the perspective of industry: the relationships between investor sentiment and industry indexes are different. (1) The lag order of the model is different. The lag orders of 10 industries such as transportation and storage, accommodation and catering, information technology, finance, real estate, scientific research technology are relatively short; (2) Granger causality is different. There is no two-way Granger causality between the four industries and investor sentiment, including hydropower and gas, accommodation and catering, leasing of business and health and society. (3) The industry indexes in the public and cyclical sectors are more affected by investor sentiment, while the industry indexes in the TMT sector are less affected by investor sentiment.Finally, this paper puts forward several suggestions, including enriching the available arbitrage tools in financial market, accelerating the development of capital market, and strengthening information disclosure.
