Search

This paper investigates how classical measurement error and additive outliers (AO) influence tests for structural change based onF-statistics. We derive theoretically the impact of general additive disturbances in the regressors on the asymptotic distribution of these tests for structural change. The small sample properties in the case of classical measurement error and AO are investigated via Monte Carlo simulations, revealing that sizes are biased upwards and that powers are reduced. Two-wavelet-based denoising methods are used to reduce these distortions. We show that these two methods can significantly improve the performance of structural break tests. [ABSTRACT FROM AUTHOR]

The hidden Markov model (HMM) has been widely used in regime classification and turning point detection for econometric series after the decisive paper by Hamilton (Econometrica 57(2):357-384, ). The present paper will show that when using HMM to detect the turning point in cyclical series, the accuracy of the detection will be influenced when the data are exposed to high volatilities or combine multiple types of cycles that have different frequency bands. Moreover, outliers will be frequently misidentified as turning points. The present paper shows that these issues can be resolved by wavelet multi-resolution analysis based methods. By providing both frequency and time resolutions, the wavelet power spectrum can identify the process dynamics at various resolution levels. We apply a Monte Carlo experiment to show that the detection accuracy of HMMs is highly improved when combined with the wavelet approach. Further simulations demonstrate the excellent accuracy of this improved HMM method relative to another two change point detection algorithms. Two empirical examples illustrate how the wavelet method can be applied to improve turning point detection in practice. [ABSTRACT FROM AUTHOR]

ABSTRACT This paper concentrates on comparing estimation and forecasting ability of quasi-maximum likelihood (QML) and support vector machines (SVM) for financial data. The financial series are fitted into a family of asymmetric power ARCH (APARCH) models. As the skewness and kurtosis are common characteristics of the financial series, a skew- t distributed innovation is assumed to model the fat tail and asymmetry. Prior research indicates that the QML estimator for the APARCH model is inefficient when the data distribution shows departure from normality, so the current paper utilizes the semi-parametric-based SVM method and shows that it is more efficient than the QML under the skewed Student's- t distributed error. As the SVM is a kernel-based technique, we further investigate its performance by applying separately a Gaussian kernel and a wavelet kernel. The results suggest that the SVM-based method generally performs better than QML for both in-sample and out-of-sample data. The outcomes also highlight the fact that the wavelet kernel outperforms the Gaussian kernel with lower forecasting error, better generation capability and more computation efficiency. Copyright © 2014 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

For testing unit root in single time series, most tests concentrate on the time domain. Recently, Fan and Gençay (Econom Theory 26:1305-1331, ) proposed a wavelet ratio test which took advantage of the information from the frequency domain by using a wavelet spectrum methodology. This test shows a better power than many time domain based unit root tests including the Dickey-Fuller (J Am Stat Assoc 74:427-431, ) type of test in the univariate time series case. On the other hand, various unit root tests in multivariate time series have appeared since the pioneering work of Levin and Lin (Unit root test in panel data: new results, University of California at San Diego, Discussion Paper, ). Among them, the Im-Pesaran-Shin (IPS) (J Econ 115(1):53-74, ) test is widely used for its straightforward implementation and robustness to heterogeneity. The IPS test is a group mean test which uses the average of the test statistics for each single series. As the test statistics in each series can be flexible, this paper will apply the wavelet ratio statistic to give a comparison with the test by using Dickey-Fuller t statistic in the single series. Simulation results show a gain in power by employing the wavelet ratio test instead of the Dickey-Fuller t statistic in the panel data case. As the IPS test is sensitive to cross sectional dependence, we further compare the robustness of both test statistics when there exists cross correctional dependence among the units in the panel data. Finally we apply a residual based wavestrapping methodology to reduce the over biased size problem brought up by the cross correlation for both test statistics. [ABSTRACT FROM AUTHOR]

In this paper, we use simulated data to investigate the power of different causality tests in a two-dimensional vector autoregressive (VAR) model. The data are presented in a nonlinear environment that is modelled using a logistic smooth transition autoregressive function. We use both linear and nonlinear causality tests to investigate the unidirection causality relationship and compare the power of these tests. The linear test is the commonly used Granger causality F test. The nonlinear test is a non-parametric test based on Baek and Brock [A general test for non-linear Granger causality: Bivariate model. Tech. Rep., Iowa State University and University of Wisconsin, Madison, WI, 1992] and Hiemstra and Jones [Testing for linear and non-linear Granger causality in the stock price–volume relation, J. Finance 49(5) (1994), pp. 1639–1664]. When implementing the nonlinear test, we use separately the original data, the linear VAR filtered residuals, and the wavelet decomposed series based on wavelet multiresolution analysis. The VAR filtered residuals and the wavelet decomposition series are used to extract the nonlinear structure of the original data. The simulation results show that the non-parametric test based on the wavelet decomposition series (which is a model-free approach) has the highest power to explore the causality relationship in nonlinear models. [ABSTRACT FROM AUTHOR]

In this article, we use the wavelet technique to improve the over-rejection problem of the traditional Dickey-Fuller tests for unit root when the data is associated with volatility like the GARCH(1, 1) effect. The logic of this technique is based on the idea that the wavelet spectrum decomposition can separate out information of different frequencies in the data series. We prove that the asymptotic distribution of the test in the wavelet environment is still the same as the traditional Dickey-Fuller type of tests. The finite sample property is improved when the data suffers from GARCH error. The investigation of the size property and the finite sample distribution of the test is carried out by Monte Carlo experiment. An empirical example with data on the net immigration to Sweden during the period 1950-2000 is used to illustrate the performance of the wavelet improved test under GARCH errors. The results reveal that using the traditional Dickey-Fuller type of tests, the unit root hypothesis is rejected while our wavelet improved test do not reject as it is more robust to GARCH errors in finite samples. [ABSTRACT FROM AUTHOR]