Asymmetry risk, state variables and stochastic discount factor specification in asset pricing models

My thesis focuses on the introduction of asymmetry in asset pricing models and portfolio selection. In the first chapter, we use a small noise expansion approach to investigate how the market equilibrium discloses, through quantities and prices, investors' preferences for three characteristics of asset returns: expected return, variance and skewness. In the second chapter, taking into account asset higher moments, we find a new bound on the volatility of any admissible stochastic discount factor (SDF) that prices correctly a set of primitive asset returns and derivatives which payoffs are a quadratic function of the same primitive assets. We further propose a method for portfolio selection which accounts for higher moments, in particular skewness. In the third chapter, we develop a utility-based economic model with state dependence in fundamentals and preferences which rationalizes and explains the risk aversion and pricing kernel puzzles put forward in Jackwerth (2000, RFS). Chapter four proposes a lattice-based model for valuing derivatives when the underlying process is affected by an unobservable state variable.; The first chapter examines how the market equilibrium discloses, through quantities and prices, investors' preferences for three characteristics of asset returns: expected return, variance and skewness. We use a small-noise expansion approach to compute heterogeneous agents' demands for several risky assets. The idea is to consider a risky asset in positive net supply which represents the market portfolio and derivatives assets in zero net supply which payoffs are nonlinear functions of the market return.; The second chapter extends the well-known Hansen and Jagannathan (1991, JPE) volatility bound. Hansen and Jagannathan characterize the volatility lower bound of any admissible SDF that prices correctly a set of primitive asset returns. We characterize this lower bound for any admissible SDF that prices correctly both primitive asset returns and quadratic payoffs of the same primitive assets. In particular, we aim at pricing derivatives which payoffs are defined as nonlinear functions of the underlying asset payoffs.; The third chapter examines the ability of economic models with regime shifts to rationalize and explain the risk aversion and pricing kernel puzzles put forward in Jackwerth (2000, RFS). We build an economy where state dependences are introduced either in investors' preferences or fundamentals and simulate European call option prices.; The last chapter presents a lattice-based method for valuing derivatives when the underlying process is affected by an unobservable state variable. (Abstract shortened by UMI.)...