Applications Of Non Linear Time Series Models On Finance And Macroeconomics

Author : Songlin Zeng
Publisher :
ISBN 13 :
Total Pages : 0 pages
Book Rating : 4.:/5 (873 download)

DOWNLOAD NOW!

Download or read book Nonlinear Time Series Models with Applications in Macroeconomics and Finance written by Songlin Zeng and published by . This book was released on 2013 with total page 0 pages. Available in PDF, EPUB and Kindle. Book excerpt: The following three chapters investigate: 1) whether Southeast Asian real exchange rates are nonlinear mean reverting, 2) bayesian inference on nonlinear time series model with applications in real exchange rate, and 3)cyclicality and bounce-back effect in stock market. Since the late nineties, both theoretical and empirical analyses devoted to the real exchange rate suggest that their dynamics might be well approximated by nonlinear models. This paper examines this possibility for post-1970 monthly ASEAN-5 data, extending the existing research in two directions. First, we use recently developed unit root tests which allow for more flexible nonlinear stationary models under the alternative than the commonly used Self-Exciting Threshold or Exponential Smooth Transition AutoRegressions. Second, while different nonlinear models survive the mis-specification tests, a Monte Carlo experiment from generalized impulse response functions is used to compare their relative relevance. Our results support the nonlinear mean-reverting hypothesis, and hence the Purchasing Power Parity, in half the cases and point to the Multiple Regime-Logistic Smooth Transition and the Self-Exciting Threshold AutoRegressive models as the most likely data generating processes of these real exchange rates.Various nonlinear threshold models are employed to mimic the real exchange rate dynamics. A natural question arises: Which model does the best job of modeling the real exchange rate process? It is difficult and not straightforward to formally compare the nonlinear models within classic approach. In the second chapter, we propose to use Bayesian approach to address this issue. The second part of my dissertation actually uses a Bayesian method to estimate some nonlinear time series models, the ACR model, SETAR model, and MAR model. We propose a full Bayesian inference approach and particular attention is paid to the parameters of the threshold variables. We discuss the choice of the prior distributions and propose a Markov-chain Monte Carlo algorithm for estimating both the parameters and the latent variables. A simulation study and the application to real exchange rate data illustrate the analysis. Our empirical results of the second chapter show that i) Bayesian estimations closely match those of the Maximum likelihood for French real exchange rate vis-a-vis Deutsche Mark; ii)the speed of real exchange rate's adjustment to equilibrium level is overestimated if heterogeneous variances in two regimes is not taken into account; iii) ACR model is preferred to other nonlinear threshold models, SETAR and MAR; iv) within ACR class models, the suitable transition function form is selected based on Bayes factor.This paper proposes an empirical study of the shape of recoveries in financial markets from a bounce-back augmented Markov Switching model. It relies on models first applied by Kim, Morley et Piger [2005] to the business cycle analysis. These models are estimated for monthly stock market returns data of five developed countries for the post-1970 period. Focusing on a potential bounce-back effect in financial markets, its presence and shape are formally tested. Our results show that i) the bounce-back effect is statistically significant and large in all countries, but Germany where evidence is less clear-cut and ii) the negative permanent impact of bear markets on the stock price index is notably reduced when the rebound is explicitly taken into account.