تمایلات سرمایه، اطلاعات و مدل قیمت گذاری دارایی

We present an asset pricing model with investor sentiment and information, which shows that the investor sentiment has a systematic and significant impact on the asset price. The equilibrium price's rational term drives the asset price to the rational, and the sentiment term leads to the asset price deviating from it. In our model, the proportion of sentiment investors and the information quality could amplify the sentiment shock on the asset price. Finally, the information is fully incorporated into prices when sentiment investors learn from prices. The model could offer a partial explanation of some financial anomalies: price bubbles, high volatility, asset prices' momentum effect and reversal effect.

The Efficient Market Hypothesis (EMH) of standard financial theory suggests that the financial market is “informationally efficient,” and rational arbitrage would eliminate irrational effect on asset prices and necessarily brings prices closer to fundamentals. However, since the 1970s, many investor abnormal behavior and financial market's anomalies, which are thought as EMH paradoxes, have begun to emerge. At the same time, behavioral asset pricing theory gradually starts to form as a complement to the traditional asset pricing theory. According to the irrational form, investors in the actual financial market may be affected by noise, cognitive biases, or investor sentiment. Some noise trader models are proposed to illustrate the influence of noise trader on asset prices (see, e.g. Black, 1986, De Long et al., 1990, Grossman and Stiglitz, 1980, Kyle, 1985 and Mendel and Shleifer, 2012); moreover, some psychology biases trader models are set up which argue that investor cognitive biases have an important effect on asset prices (e.g. Barberis et al., 1998, Daniel et al., 1998, Hong and Stein, 1999 and Yan, 2010). The shortcomings of both types of models are that the noise and psychology biases are difficult to identify and can't be measured in the securities market; consequently, they can't be empirically testified. Compared with the noise term and bias factor in the securities market, investor sentiment could be quantitatively measured; furthermore, the corresponding empirical analysis can be made (Baker and Wurgler, 2006 and Baker and Wurgler, 2007). In recent years, the systematic role of investor sentiment has been investigated by many empirical analyses and theoretical studies. Some empirical results show that investor sentiment has an important and systematic effect on asset pricing (Baker and Wurgler, 2006, Baker and Wurgler, 2007, Baker et al., 2012, Brown and Cliff, 2004, Brown and Cliff, 2005, Kumar and Lee, 2006, Lee et al., 2002, Seybert and Yang, 2012, Stambaugh et al., 2012 and Yu and Yuan, 2011). However, the sentiment-based asset pricing model is still in the exploratory stage. Some sentiment asset pricing models have been presented to emphasize the systematic role of investor sentiment in asset pricing. For instance, Yang and Yan (2011) set up a sentiment asset pricing model with representative sentiment investors, Yang et al. (2012) propose a sentiment capital asset pricing model and the result shows that different investor sentiments lead to different perceived prices, and Yang and Zhang, 2013a and Yang and Zhang, 2013b consider a sentiment asset pricing model with consumption. Nevertheless, the related sentiment asset pricing models don't possess the generality of analysis, which only focus on investor sentiment and don't involve the important factors such as fundamental information. Much different from the previous literature on sentiment asset pricing model, we present a generalized sentiment asset pricing model with information based on the framework of Grossman and Stiglitz (1980). We consider one class of uninformed sentiment investors who are vulnerable to sentiment and trade on it, so our model focuses on the interaction of rational investors and uninformed sentiment investors and shows how this interaction could sustain incorrect prices. Ultimately, we demonstrate how the financial asset is priced when sentiment investors learn from prices. The features of our model, which distinguish it from the previous sentiment asset pricing models, are the following terms. First, it gives an analytical solution to the sentiment equilibrium price which could be decomposed to the rational term and the sentiment term, the equilibrium price's rational term makes the asset price return to the rational expected value, and the sentiment term leads to the asset price deviating from the rational expected value which can generate price bubbles and high volatility. Second, when sentiment investors' proportion is less than a constant value, price move in reaction to the arrival of information is on average positively correlated with later price change and the changes of asset prices show short-term momentum effect; however, when sentiment investors' proportion is more than a constant value, price move resulting from information arrival is on average negatively correlated with later price change and the changes of asset prices show long-term reversal effect. Third, increasing the proportion of sentiment investors, and decreasing the quality of information would increase the sensitivity coefficient on sentiment; on the contrary, it is to the sensitivity coefficient on information. Fourth, adding more rational investors, increasing the information quality, and decreasing the sentiment expansion coefficient would increase the informativeness of the price system and the market efficiency. Finally, when uninformed sentiment investors learn from prices, all the information is incorporated into prices; increasing the proportion of rational investors, and decreasing sentiment expansion coefficient would increase the market efficiency, but the quality of information no longer has an effect in this case. The rest of the paper is organized as follows. In Section 2, we spell out the economy for formal model. In Section 3, we consider the benchmark case in which investors are homogenous. In Section 4, we present a generalized sentiment asset pricing model with information. Section 5 concludes.

Many abnormal investor behavior and financial market's anomalies can't be well explained by the standard financial theory, some scholars have begun to accept the argument that market is not always effective, and have presented behavioral asset pricing theory based on bounded rationality and limited arbitrage. In recent years, some empirical studies have shown that investor sentiment has a systematic and significant impact on the financial asset price. Nevertheless, the sentiment-based asset pricing model is still in the exploratory stage. Some sentiment asset pricing models have been developed to emphasize the systematic role of investor sentiment in asset prices. However, these sentiment asset pricing models have many shortages and do not possess the generality of analysis. We present a generalized sentiment asset pricing model with information based on the framework of Grossman and Stiglitz (1980). Summaries and conclusions about the characteristics of our model are as follows: Firstly, an analytical solution to the sentiment equilibrium price is given, it could be decomposed to the rational term and the sentiment term and offers a partial explanation for the financial anomalies of price bubbles and high volatility. Secondly, when sentiment investors' proportion is less than a constant value, the changes of asset prices show short-term momentum effect; however, when sentiment investors' proportion is more than a constant value, the changes of asset prices show long-term reversal effect. Thirdly, adding more rational investors, increasing the information quality, decreasing the sentiment expansion coefficient, decreasing sentiment shock, and increasing information shock for a constant n would increase the informativeness of the price system and the market efficiency. Finally, when uninformed sentiment investors learn from prices, all the information is incorporated into prices, and the quality of information no longer has an effect in this case. Our findings could raise some interesting issues for future research. For example, in order to analyze the dynamic effects of investor sentiment on asset pricing with information, one needs to build a dynamic sentiment asset pricing model, and even a continuous sentiment asset pricing model with information.