Measuring The Risk Aversion Based On Single Variable GARCH-M Model

In this thesis, we mainly study the problem of measuring the risk aversion ofthe stock market, aim at finding a more appropriate way to estimate the aggregaterisk aversion of the stock market, especially in the situation of a global ficial marketwith a fierce volatility. To begin with, in the exordium, we introduce the concepts ofindividual risk aversion and the aggregate risk aversion. In chapter 2, we illustrate theconstant-parametric ARCH-M model, the time-varying parameter ARCH-M model,univariate functional coefficient ARCH-M model and adaptive functional coefficientARCH-M model, and we also simplly analyse those models.In the subsequent chapter 3, we propose GARCH-M model to measure the riskaversion. In this model, the risk aversion is regarded as an unknown function of someeconomic variables.We first estimate this function by the way of profile likelihood,andthen estimate the unknown parameters by QMLE. Asymptotic properties of the esti-mators are discussed later. We also give the ways to choose the window widths.In chapter 4, we do some simulations about the proposed model and find that ourmethod performs well. Then we apply the proposed model to analyze American stockmarket based on three macroeconomic elements to measure risk aversion and explainthe estimated results.In chapter 5, we extend the proposed model in chapter 3 and discuss the potentialproblems and where need to be improved.