Theory And Positive Research On Modern Portfolio

Modern portfolio theory (MPT) provides the foundation for modern finance theory and investment theory. It explores how risk-averse investors construct portfolios in order to optimize expected utility for a given level of market risk. This thesis investigates MPT both theoretically and practically. It consists of five chapters.Chapter one introduces how to construct the mean-variance model based on Markowitz's portfolio theory. Furthermore, the efficient frontier is derived, and the mean-variance model is extended to relax the constraints. To avoid the disadvantages of variance in measuring market risk, the semi-variance and LPM methods are introduced. The three measures, variance, semi-variance, and LPM, are compared.Chapter two analyzes the capital assets pricing model (CAPM) and discusses how assets are priced when the market is in equilibrium. Based on the discussion, we relax some constraints in the traditional CAPM. In particular, we investigate three variations of the traditional CAPM, which are the CAPM with riskless assets, the CAPM with non-marketable assets, and the consumption capital asset pricing model. Finally, with the help of the modern behavior finance theory, the pricing model with irrational expectations is introduced.Chapter three analyzes the Arbitrage Pricing Theory (APT) of risk assets. How to choose the factor is one core problem in APT. We go deep into this problem, and two popular methods are introduced. We analyze their applicability and point out their advantage and disadvantage respectively.Chapter four analyzes VaR, a new tool for measuring the market risk and examines the four main methods to calculate VaR. The historical simulation method, which is being widely used in risk management area, has a shortage of neglecting many information when calculated only by close price series. In this chapter we present a new historical simulation method which also takes into account the open price, daily high price, and daily low price so as to improve the calculation method of VaR.Chapter five is positive research. Refer to cross-sectional regression model introduced by Fama and Macbeth, we uses two-moment regression method to examine Shenzhen B Stock Market. The positive result displays that the pricing behavior of Shenzhen B stock market is different with CAPM and non-systemic risk can explain stock profit in a certain extent.