Customer Reviews

Financial Modeling Under Non-Gaussian Distributions (Springer Finance)

5 star:		(1)
4 star:		(0)
3 star:		(0)
2 star:		(0)
1 star:		(0)

 

 

This book is an outstanding a clear presentation of non-Gaussian financial modeling. In financial markets, the Gaussian curve or bell curve, is not accurate in that most markets are skewed (a predisposition to grow on average, not zero) and fat-tailed (rare events such as market crashes happen more often than a Gaussian curve would suggest). Therefore, non-Gaussian modeling is essential to make money in the market or assess risk. This book goes through all the new techniques of non-Gaussian modeling. It does an exceptional job discussing the GARCH generalized autoregressive conditional heteroskedasticity. This is but a fancy word for fluctuations in volatility over time pretty much dependent on recent fluctuations. It works very well I must say empirically, and has tripled the rationality and profitability of my portfolio - especially one of the versions of the GARCH over the others reviewed - but which one I'd rather not say, for obvious reasons ;) The book is weakest at page 183 or so, with the models and I was rather disappointed with the exclusion of the market crash of the 80s in the empirical analysis - wouldn't rare events be the main reason for improving non-Gaussian modeling? Anyway it's rather poor, but thorough, with additive and multivariate GARCHes but the fault lies with the faultiness of the theories not the authors, at least they're encyclopedic. The book picks up at the end with copulas, and a complete discussion of non-Gaussian option pricing. The review of BSM is appreciated and actually well-done, and a nice reminder of what we are trying to improve on exactly. I think this is a most incredible book, very clearly written, and at times, quite an enjoyable read for such a topic. All it takes is multivariate calculus and basic statistics, but more math ability will make the implications and comments breathtaking at times. I often find myself inspired by a passage or footnote to create a whole subroutine in R or python. I think avoiding Bayesian topics and Monte Carlo was disappointing, but wise in terms of focus. A great book for graduate mathematics in applications of statistics or stochastic calculus, or a good book for modeling fundamentals in economics or business management at the post-graduate level.