Financial modeling under non-Gaussian distributions. (Q855067): Difference between revisions

This book is written for non-mathematicans who want to model financial market prices. This book does not sacrify the mathematical rigor and the complexity of the original models. It targets practioners in the financial industry. It is suitable for use as core text for students in empirical finance, financial econometrics and financial derivatives. It is useful for mathematician who want to know more about their mathematical tools are applied in finance. The reader should have some basic knowledge in statistics, calculus, and probability theory. This book has five parts. Part I describes financial markets and financial time series. Part II treats econometric modeling of asset returns. Part III presents applications of non-Gaussian econometrics. Option pricing with non-Gaussian returns are treated in Part IV. Appendices on option pricing mathematics finish the book in part V. Part I has three chapters. After chapter I (introduction), chapter 2 discusses the unique statistical properties and financial market data and several so-called stylized facts being the basis of Part II. Chapter 3 describes the actual functioning and microstructure of financial markets. Some theoretical models are presented. They help to explain why asset returns are non-normal and time-dependent. Part II (chapters 4--7) is concerned with the time series aspects of asset returns. Chapters 4, 5, and 7 cover models for the second, third, and fourth moments and the tails of return distributions. Chapter 6 deals with the dependence structure when the higher moments display significant departure from normality and returns appear to be time dependent. Chapters 4 and 5 are concerned with univariate time series characteristics. Chapter 6 shifts the focus to the relationship between and among the asset returns series. Chapter 7 deals with models for only the tails of the distribution. It describes various approaches for characterizing the behavior of extreme points. Part III (chapters 8, 9) presents some applications of the models described in Part I. Chapter~8 deals with risk management and the Value-at-Risk measure. Chapter 9 is concerned with portfolio construction and asset allocation. Markowitz's mean-variance analysis may not hold any more in the context of non-normal distributions. The main idea is that the investor's expected utility may be approximated as a function of mean, variance, but also of higher moments of portfolio return. Part IV (chapters 10--12) deals with derivative assets and option pricing. Chapter 10 goes through the fundamentals of the Black-Scholes-Merton model and uses it as an example to introduce the key mathematical concepts as Brownian motion and stochastic calculus (chapter 13) and martingales and changing measure (chapter 14). It is based on the underlying asset return having a normal distribution. From the Black-Scholes-Merton model it emerges that options can be priced using risk neutral densities (RNDs). The most important fact is that these empirically obtained RNDs are almost exclusively non-Gaussian. Chapter 11 covers a whole range of parametric and nonparametric methods for extracting RNDs. These techniques do not assume a specific model for the underlying asset. Chapter 12 puts extensions of the Black-Scholes-Merton model into the ``structural'' option pricing model. This chapter is the most mathematically demanding and would require chapters 15, 16, and 17 of Part V. But this chapter reflects the non-Gaussian nature of the underlying asset distributions. Part V (chapter 13--17) treats the mathematical background as Brownian motion and stochastic calculus (chapter 13), martingale and changing measure (chapter 14), characteristic functions and Fourier transforms (chapter 15), jump processes (chapter 16), and Lévy processes (chapter~17).