nep-ets New Economics Papers
on Econometric Time Series
Issue of 2012‒02‒01
six papers chosen by
Yong Yin
SUNY at Buffalo

  1. A Moment-Matching Method for Approximating Vector Autoregressive Processes by Finite-State Markov Chains By Nikolay Gospodinov; Damba Lkhagvasuren
  2. The Wavelet-based Estimation for Long Memory Signal Plus Noise Models By Kei Nanamiya
  3. Regression with a Slowly Varying Regressor in the Presence of a Unit Root By Yoshimasa Uematsu
  4. Nonparametric LAD Cointegrating Regression By Toshio Honda
  5. A Nonlinear Panel Data Model of Cross-Sectional Dependence By James Mitchell; George Kapetanios; Yongcheol Shin
  6. Estimating VAR's sampled at mixed or irregular spaced frequencies : a Bayesian approach By Ching Wai (Jeremy) Chiu; Bjørn Eraker; Andrew T. Foerster; Tae Bong Kim; Hernán D. Seoane

  1. By: Nikolay Gospodinov (Concordia University and CIREQ); Damba Lkhagvasuren (Concordia University and CIREQ)
    Abstract: This paper proposes a moment-matching method for approximating vector autoregressions by finite-state Markov chains. The Markov chain is constructed by targeting the conditional moments of the underlying continuous process. The proposed method is more robust to the number of discrete values and tends to outperform the existing methods for approximating multivariate processes over a wide range of the parameter space, especially for highly persistent vector autoregressions with roots near the unit circle.
    Keywords: Markov Chain, Vector Autoregressive Processes, Functional Equation, Numerical Methods, Moment Matching
    JEL: C15 C60
    Date: 2011–06–08
  2. By: Kei Nanamiya
    Abstract: We propose new wavelet-based procedure to estimate the memory parameter of an unobserved process from an observed process affected by noise in order to improve the performance of the estimator by taking into account the dependency of the wavelet coefficients of long memory processes. In our procedure, using the AR (1) approximation for the wavelet transformed long memory processes which is introduced by Craigmile, Guttorp and Percival (2005), we apply the ARMA (1, 1) approximation to the wavelet coefficients of the observed process at each level. We also compare this procedure to the usual wavelet-based procedure by numerical simulations.
    Keywords: Wavelet, Long Memory Process, Measurement Error Problem
    Date: 2011–12
  3. By: Yoshimasa Uematsu
    Abstract: This paper considers the regression model with a slowly varying (SV) regressor in the presence of a unit root in serially correlated disturbances. This regressor is known to be asymptotically collinear with the constant term; see Phillips (2007). Under nonstationarity, we find that the estimated coefficients of the constant term and the SV regressor are asymptotically normal, but neither is consistent. Further, we derive the limiting distribution of the unit root test statistic. We may here observe that the finite sample approximation to the limiting one is not monotone and it is poor due to the influence of the collinear regressor. In order to construct a well-behaved test statistic, we recommend dropping the constant term intentionally from the regression and computing the statistics, which are still consistent under the true model having the constant term. The powers and sizes of these statistics are found to be well-behaved through simulation studies. Finally, these results are extended to general Phillips and Perron-type statistics.
    Date: 2011–10
  4. By: Toshio Honda
    Abstract: We deal with nonparametric estimation in a nonlinear cointegration model whose regressor and dependent variable can be contemporaneously correlated. The asymptotic properties of the Nadaraya-Watson estimator are already examined in the literature. In this paper, we consider nonparametric least absolute deviation (LAD) regression and derive the asymptotic distributions of the local constant and local linear estimators by appealing to the local time approach.
    Keywords: Nonlinear Cointegration, Integrated Process, Local Time, Least Absolute Deviation, Local Polynomial Regression, Bias
    Date: 2011–10
  5. By: James Mitchell; George Kapetanios; Yongcheol Shin
    Abstract: This paper proposes a nonlinear panel data model which can generate endogenously both `weak' and `strong' cross-sectional dependence. The model's distinguishing characteristic is that a given agent's behaviour is influenced by an aggregation of the views or actions of those around them. The model allows for considerable flexibility in terms of the genesis of this herding or clustering type behaviour. At an econometric level, the model is shown to nest various extant dynamic panel data models. These include panel AR models, spatial models, which accommodate weak dependence only, and panel models where cross-sectional averages or factors exogenously generate strong, but not weak, cross sectional dependence. An important implication is that the appropriate model for the aggregate series becomes intrinsically nonlinear, due to the clustering behaviour, and thus requires the disaggregates to be simultaneously considered with the aggregate. We provide the associated asymptotic theory for estimation and inference. This is supplemented with Monte Carlo studies and two empirical applications which indicate the utility of our proposed model as both a structural and reduced form vehicle to model different types of cross-sectional dependence, including evolving clusters.
    Keywords: Nonlinear Panel Data Model; Clustering; Cross-section Dependence; Factor Models; Monte Carlo Simulations; Application to Stock Returns and Inflation Expectations
    JEL: C31 C33 C51 E31 G14
    Date: 2012–01
  6. By: Ching Wai (Jeremy) Chiu; Bjørn Eraker; Andrew T. Foerster; Tae Bong Kim; Hernán D. Seoane
    Abstract: Economic data are collected at various frequencies but econometric estimation typically uses the coarsest frequency. This paper develops a Gibbs sampler for estimating VAR models with mixed and irregularly sampled data. The approach allows efficient likelihood inference even with irregular and mixed frequency data. The Gibbs sampler uses simple conjugate posteriors even in high dimensional parameter spaces, avoiding a non-Gaussian likelihood surface even when the Kalman filter applies. Two applications illustrate the methodology and demonstrate efficiency gains from the mixed frequency estimator: one constructs quarterly GDP estimates from monthly data, the second uses weekly financial data to inform monthly output.
    Date: 2011

This nep-ets issue is ©2012 by Yong Yin. It is provided as is without any express or implied warranty. It may be freely redistributed in whole or in part for any purpose. If distributed in part, please include this notice.
General information on the NEP project can be found at For comments please write to the director of NEP, Marco Novarese at <>. Put “NEP” in the subject, otherwise your mail may be rejected.
NEP’s infrastructure is sponsored by the School of Economics and Finance of Massey University in New Zealand.