
on Econometrics 
By:  JeanMarie Dufour; Abdeljelil Farhat; Marc Hallin 
Abstract:  We consider the problem of testing whether the observations X1, · · ·, Xn of a time series are independent with unspecified (possibly nonidentical) distributions symmetric about a common known median. Various bounds on the distributions of serial correlation coefficients are proposed: exponential bounds, Eatontype bounds, Chebyshev bounds and BerryEsséenZolotarev bounds. The bounds are exact in finite samples, distributionfree and easy to compute. The performance of the bounds is evaluated and compared with traditional serial dependence tests in a simulation experiment. The procedures proposed are applied to U.S. data on interest rates (commercial paper rate). <P>Nous étudions le problème qui consiste à tester l’hypothèse que des observations X1, · · ·, Xn d’une série chronologique sont indépendantes avec des distributions non spécifiées (possiblement non identiques) symétriques autour d’une médiane connue. Nous proposons plusieurs bornes sur les distributions des coefficients d’autocorrélation : bornes exponentielles, bornes de type Eaton, bornes de Chebyshev et bornes de BerryEsséenZolotarev. Les bornes sont exactes dans les échantillons finis, non paramétriques et faciles à calculer. Nous évaluons par simulation la performance des bornes et comparons celleci à celle de tests d’autocorrélation traditionnels. Les procédures proposées sont appliquées à des données de taux d’intérêt américaines (“commercial paper rate”). 
Keywords:  autocorrelation; serial dependence; nonparametric test; distributionfree test; heterogeneity; heteroskedasticity; symmetric distribution; robustness; exact test; bound; exponential bound; large deviations; Chebyshev inequality; BerryEsséen; interest rates, autocorrelation; serial dependence; nonparametric test; distributionfree test; heterogeneity; heteroskedasticity; symmetric distribution; robustness; exact test; bound; exponential bound; large deviations; Chebyshev inequality; BerryEsséen; interest rates 
JEL:  C14 C22 C12 C32 E4 
Date:  2005–02–01 
URL:  http://d.repec.org/n?u=RePEc:cir:cirwor:2005s04&r=ecm 
By:  Cizek,P. (Tilburg University, Center for Economic Research) 
Abstract:  High breakdownpoint regression estimators protect against large errors and data contamination. Motivated by some { the least trimmed squares and maximum trimmed likelihood estimators { we propose a general trimmed estimator, which uni¯es and extends many existing robust procedures. We derive here the consistency and rate of convergence of the proposed general trimmed estimator under mild ¯mixing conditions and demonstrate its applicability in nonlinear regression, time series, limited dependent variable models, and panel data. 
JEL:  C13 C20 C24 C25 
Date:  2004 
URL:  http://d.repec.org/n?u=RePEc:dgr:kubcen:2004130&r=ecm 
By:  A.H.J. den Reijer 
Abstract:  This paper applies large scale factor models to Dutch quarterly data inorder to generate forecasts of GDP growth rates for an horizon up to 8 quarters ahead. The data set consists of the series underlying the cen tral bank´s macroeconomic structural model for the Netherlands sup plemented with leading indicator variables. In a pseudo outofsample forecasting context, we select optimal models in the time dimension and the optimal size of the ordered data set in the crosssectional dimension. The main empirical ?ndings of this paper are that the crosssectional opti mization substantially improves the forecasting performance of the factor models. However, only the dynamic factor model systematically outper forms and encompasses the autoregressive benchmark model with an op timal subset of the data of around 110 series. The forecasting gains in terms of mean squared errors range from 10% to 30% for forecast horizons up to 6 quarters ahead. 
Keywords:  Factor models; Forecasting; Leading Indicators. 
JEL:  C43 C51 E32 
Date:  2005–02 
URL:  http://d.repec.org/n?u=RePEc:dnb:dnbwpp:028&r=ecm 
By:  Hiroaki Chigira 
Date:  2005–01 
URL:  http://d.repec.org/n?u=RePEc:hst:hstdps:d0469&r=ecm 
By:  Wiliam Branch (University of Californis  Irvine); George W. Evans (University of Oregon Economics Department) 
Abstract:  We compare the performance of alternative recursive forecasting models. A simple constant gain algorithm, used widely in the learning literature, both forecasts well out of sample and also provides the best fit to the Survey of Professional Forecasters. 
Keywords:  constant gain, recursive learning, expectations 
JEL:  E37 D84 D83 
Date:  2005–02–01 
URL:  http://d.repec.org/n?u=RePEc:ore:uoecwp:20053&r=ecm 
By:  Stephen G. Donald (University of Texas at Austin); Natércia Fortuna (CEMPRE, Faculdade de Economia do Porto); Vladas Pipiras (University of North Carolina at Chapel Hill) 
Abstract:  We focus on the problem of rank estimation in an unknown symmetric matrix based on a symmetric, asymptotically normal estimator of the matrix. The related positive definite limit covariance matrix is assumed to be estimated consistently, and to have either a Kronecker product or an arbitrary structure. These assumptions are standard although they also exclude the case when the matrix estimator is positive or negative semidefinite. We adapt and reexamine here some available rank tests, and introduce a new rank test based on the eigenvalues of the matrix estimator. We discuss several applications where rank estimation in symmetric matrices is of interest, and also provide a small simulation study and an application. 
Keywords:  rank, symmetric matrix, eigenvalues, matrix decompositions, estimation, asymptotic normality, consistency 
JEL:  C12 C13 
Date:  2005–02 
URL:  http://d.repec.org/n?u=RePEc:por:fepwps:167&r=ecm 
By:  Yoichi Arai (Faculty of Economics, University of Tokyo); Eiji Kurozumi (Department of Economics, Hitotsubashi University) 
Abstract:  In this paper we propose residualbased tests for the null hypothesis of cointegration with structural breaks against the alternative of no cointegration. The Lagrange Multiplier test is proposed and its limiting distribution is obtained for the case in which the timing of a structural break is known. Then the test statistic is extended in two ways to deal with a structural break of unknown timing. The first test statistic, a plugin version of the test statistic for known timing, replaces the true break point by the estimated one. We also propose a second test statistic where the break point is chosen to be most favorable for the null hypothesis. We show the limiting properties of both statistics under the null as well as the alternative. Critical values are calculated for the tests by simulation methods. Finitesample simulations show that the empirical size of the test is close to the nominal one unless the regression error is very persistent and that the test rejects the null when no cointegrating relationship with a structural break is present. 
Date:  2005–02 
URL:  http://d.repec.org/n?u=RePEc:tky:fseres:2005cf319&r=ecm 
By:  Henry Kim 
Abstract:  We describe an algorithm for calculating second order approximations to the solutions to nonlinear stochastic rational expectations models. The paper also explains methods for using such an approximate solution to generate forecasts, simulated time paths for the model, and evaluations of expected welfare differences across different versions of a model. The paper gives conditions for local validity of the approximation that allow for disturbance distributions with unbounded support and allow for nonstationarity of the solution process. 
URL:  http://d.repec.org/n?u=RePEc:tuf:tuftec:0505&r=ecm 