Essays on behavioral heterogeneity and asset pricing

• Main Author: Li, Changtai
• Other Authors: Chia Wai Mun
• Format: Thesis-Doctor of Philosophy
• Language: English
• Published: Nanyang Technological University 2020
• Subjects: Social sciences::Economic development; Business::Finance::Assets
• Online Access: https://hdl.handle.net/10356/145290
• Institution: Nanyang Technological University

Recent empirical studies have confirmed the importance of investor behavior in asset pricing. This thesis aims to explore the effect of behavioral heterogeneity on asset pricing with focuses on investor sentiment and trading volume. Other related issues on financial markets, including stylized facts, fundamental value, steady states, chart patterns and the formation of sudden, smooth and disturbing crises are also investigated. I mainly explore these issues by using a heterogeneous agent model methodology. In recent years, asset pricing models based on Efficient Market Hypothesis (EMH) fail to explain several ubiquitous financial regularities, such as fat tail of returns, excess volatility and systemic under- or over-valuation of stock prices relative to their intrinsic values. It gives rise to the alternative behavioral finance theory which aims to provide rationale behind the unexplained market anomalies. Asset pricing models in behavioral finance usually incorporate the psychological factors into the decision making process and investigate the role of these factors in various market dynamics. As one of the representative factor in psychology, sentiment has attracted much attention from both academia and industry. Empirical evidences have been found that sentiment is highly related to assets returns and many market dynamics. Investors have attempted to construct sentiment index and develop sentiment-based strategies to invest in financial markets. To better understand the role of sentiment in asset pricing, a solid behavioral finance model should be built. Among the new behavioral approaches, heterogeneous agent model (HAM) is one of the best to fit the financial data and explain the market stylized facts. Most of the HAMs include two important classes of investors: fundamentalists and chartists. Fundamentalists make trading decision based upon market fundamentals, such as dividends, earnings, macroeconomic factors etc. They believe that irrational price fluctuation is temporary and price is mean-reverting to fundamental level. In contrast, chartists look for simple patterns, e.g. trends, in past prices and base their investment decisions upon extrapolation of these patterns. More complicated HAMs consider the interaction of different agents, and agents could switch their trading strategies. HAMs with interacting agents have the ability to explain most of the stylized facts observed in financial time series. Combining with sentiment, HAMs could be more powerful to demonstrate the role of investor’s behavior in asset pricing. This thesis will mainly focus on four questions: (1) How to model sentiment in HAM and what is the format of market equilibrium steady state with sentiment? (2) What is role of sentiment in stylized facts and different types of crisis in financial market? (3) How to estimate the two-market HAM with sentiment? (4) How to use HAMs to explain another important indicator, trading volume? To answer the first question, I develop a heterogeneous agent model in Chapter 2, and a sentiment index is constructed by taking into consideration of sentiment memory, social interaction of investors and sentiment shocks. From the steady state analysis of the market equilibrium, I find that sentiment models could generate both fundamental steady states featured with neutral sentiment and non-fundamental steady states with polarized sentiment. Chapter 3 applies the three-agent sentiment model proposed in Chapter 2 to conduct numerical analysis. I find the sentiment model is more powerful to explain the stylized facts in market comparing with model without sentiment. By further exploring the role of sentiment in the crisis formation, I conclude that investor sentiment could contribute to the frequency and magnitude of financial crisis. Chapter 4 examines a two-market sentiment HAM with a focus on estimation. I attempt to estimate two-market model by Bayesian approach with extended Kalman filter. Estimation results from US and UK data show the evidence of market switching behavior of internationalist group, which provides a channel to explain crisis contagion through sentiment-induced market switching of international traders. Jumping out from sentiment, Chapter 5 investigates another important indicator, trading volume, under a three-agent HAM. The model is proven to have capability of explaining the stylized facts on both market prices and trading volumes. It also provides an option to theoretically explore the co-evolvement of prices, volumes and beliefs in financial markets. The sentiment models and volume model developed in this thesis provide new angles to understand the financial markets. Both sentiment-based factors and market-based factors are important for asset pricing. With the support of these theoretical studies, one can build relevant trading strategies and investment portfolios in practice.