Papers by Matei Demetrescu
Sometimes, integration tests are applied to seasonal data without al- lowing for seasonal determi... more Sometimes, integration tests are applied to seasonal data without al- lowing for seasonal deterministics. This paper studies the efiect of neglecting seasonal dummy variables. For the Dickey-Fuller test, it is observed that the distribution is shifted to the left, with lower dis- persion at the same time, whenever deterministic seasonality is not accounted for. When accounting for serial correlation, the
The asymptotically normal LM integration test is adapted for small panels with correlated units. ... more The asymptotically normal LM integration test is adapted for small panels with correlated units. First, some conditions are discussed under which the test statistic is asymptotically normal in a single unit, or ´ 2 (1) upon squaring. Second, in the context of cross-correlated panels, we propose to rotate the panel in such a way that the time series are uncorrelated.
This paper describes the general principles and methods of the Þducial inference. A brief survey ... more This paper describes the general principles and methods of the Þducial inference. A brief survey of its competing inferencial theories as well as a comparison with them are also provided. Arguments in favour of the application of the Þducial method to the parameters of discrete random variables are given, and, as an applica- tion, the Þducial distribution associated to the
The flnite-sample behavior of the Lagrange Multiplier (LM) test for fractional inte- gration prop... more The flnite-sample behavior of the Lagrange Multiplier (LM) test for fractional inte- gration proposed by Breitung and Hassler (JoE 2002) is examined. It is found by extensive Monte Carlo simulations that size distortions can be quite large in small samples. These are caused by a flnite-sample bias towards the alternative. Analytic expressions for this bias are derived, based on which the test can easily be corrected.
A new stationarity test for heterogeneous panel data with large cross- sectional dimension is dev... more A new stationarity test for heterogeneous panel data with large cross- sectional dimension is developed and used to examine a panel with growth rates of unit labor cost in the US. The test allows for strong cross-unit dependence in the form of unbounded long-run correlation matrices, for which a simple parameterization is proposed. A KPSS- type distribution results asymptotically if letting T ! 1 be followed by N ! 1. Some evidence against stationarity (short memory) is found for the examined series.
Studying annual growth rates (seasonal differences) in case of seasonal data produces much more p... more Studying annual growth rates (seasonal differences) in case of seasonal data produces much more persistence, autocorrelation and stronger evidence in favour of a unit root than analyzing seasonal growth rates (ordinary differences). First, this statement is quantified theoretically. Second, it is supported experimentally which simulations, and, finally, it is empirically illustrated with quarterly GDP deflators from 7 European economies.
The problem of approximating unknown functions is discussed for the case where approximation erro... more The problem of approximating unknown functions is discussed for the case where approximation errors are evaluated by means of a gen- eral cost-of-error (loss) function, not necessarily the squared-error one. To ensure minimal overall loss, approximation is carried out by fltting under the relevant loss function. Convergence results are provided in a general framework, allowing among others for Taylor approximations, approxima- tions with Hermite polynomials or approximations with Neural Networks and for random measurement error.
Econometric Theory
An integration test against fractional alternatives is suggested for univariate time series. The ... more An integration test against fractional alternatives is suggested for univariate time series. The new test is a completely regression-based, lag augmented version of the Lagrange multiplier (LM) test by Robinson (1991, Journal of Econometrics 47, 67 84). Our main contributions, however, are the following. First, we let the short memory component follow a general linear process. Second, the innovations driving this process are martingale differences with eventual conditional heteroskedasticity that is accounted for by means of White s standard errors. Third, we assume the number of lags to grow with the sample size, thus approximating the general linear process. Under these assumptions, limiting normality of the test statistic is retained. The usefulness of the asymptotic results for finite samples is established in Monte Carlo experiments. In particular, several strategies of model selection are studied.An earlier version of this paper was presented at the URCT Conference in Faro, Po...
Journal of Time Series Analysis
The distributions of cointegration tests are affected when the innovation variance varies over ti... more The distributions of cointegration tests are affected when the innovation variance varies over time. In panels, one must also pay attention to dependence among units. To obtain a panel cointegration test robust to both heteroskedasticity and dependence, we adapt the nonlinear instruments method proposed for the Dickey–Fuller test by Chang (2002, J Econometrics 110, 261–292) to an error-correction framework. We show that IV-based testing of the no error-correction null in individual equations yields standard normal test statistics when computed with heteroskedasticity-robust standard errors. The result holds under endogenous regressors, irrespective of the number of integrated covariates and for any variance profile. A non-cointegration test combining single-equation tests retains these nice properties. In panels of fixed cross-sectional dimension, such test statistics from individual units are shown to be asymptotically independent even under dependence, leading to panel tests robus...
Journal of Applied Statistics
A new stationarity test for heterogeneous panel data with large cross-sectional dimension is deve... more A new stationarity test for heterogeneous panel data with large cross-sectional dimension is developed and used to examine a panel with growth rates of unit labor cost in the USA. The test allows for strong cross-unit dependence in the form of unbounded long-run correlation matrices, for which a simple parameterization is proposed. A KPSS-type distribution results asymptotically if letting T→∞ be followed by N→∞. Some evidence against stationarity (short memory) is found for the examined series.
Optimal prediction, as well as estimation, of general linear processes under asymmetric loss func... more Optimal prediction, as well as estimation, of general linear processes under asymmetric loss functions are addressed. Conditions under which autore- gressive approximations of order growing to infinity may be used to this purpose are studied for two variants of estimation under the relevant loss function. Monte Carlo simulations illustrate the asymptotic results.
The basic concepts of flducial inference are presented in the flrst part of this paper. Problems ... more The basic concepts of flducial inference are presented in the flrst part of this paper. Problems arise if the form of the distribution of the pivot used by the flducial argument is not known. Our main result consists in showing how bootstrap techniques can be used to handle this kind of situation. Keywords
Economic Quality Control, 2006
When evaluating point estimators by means of general loss functions, the expected loss is not alw... more When evaluating point estimators by means of general loss functions, the expected loss is not always minimal, similar to the case of mean-biased estimators, whose mean squared error can be reduced by accounting for the mean-bias. Depending on the loss function, the socalled Lehmann-bias can be significantly more important than the mean-bias of an estimator. Although a simple decomposition does not hold for expected losses as it does for the mean squared error, the expected loss can still be reduced by correcting for the Lehmann-bias. An asymptotic and a bootstrap-based correction are suggested and compared in small samples for the exponential distribution by means of Monte Carlo simulation.
Economic Quality Control, 2005
A procedure is derived for determining the values of the parameters of a multi- nomial distributi... more A procedure is derived for determining the values of the parameters of a multi- nomial distribution by means of the fiducial approach proposed by Fisher. By decomposing the involved probability (mass or density) functions, it is shown that the resulting fiducial distributions belong to the Dirichlet family.
Statistical Papers, 2007
Statistics and Econometric Methods, Goethe-University Frankfurt, Gr~ifstr. 78, 60054 Frankfurt, G... more Statistics and Econometric Methods, Goethe-University Frankfurt, Gr~ifstr. 78, 60054 Frankfurt, Germany; (e-mail: [email protected]) ... Abstract Whenever deterministic seasonality is ignored, the distribution ... Dickey-Fuller test without seasonal dummies is ...
Statistical Papers, 2013
ABSTRACT The paper examines the behavior of a generalized version of the nonlinear IV unit root t... more ABSTRACT The paper examines the behavior of a generalized version of the nonlinear IV unit root test proposed by Chang (2002) when the series’ errors exhibit nonstationary volatility. The leading case of such nonstationary volatility concerns structural breaks in the error variance. We show that the generalized test is not robust to variance changes in general, and illustrate the extent of the resulting size distortions in finite samples. More importantly, we show that pivotality is recovered when using Eicker-White heteroskedasticity-consistent standard errors. This contrasts with the case of Dickey-Fuller unit root tests, for which Eicker-White standard errors do not produce robustness and thus require computationally costly corrections such as the (wild) bootstrap or estimation of the so-called variance profile. The pivotal versions of the generalized IV tests – with or without the correct standard errors – do however have no power in $1/T$ -neighbourhoods of the null. We also study the validity of panel versions of the tests considered here.
Oxford Bulletin of Economics and Statistics, 2006
The inverse normal method, which is used to combine P-values from a series of statistical tests, ... more The inverse normal method, which is used to combine P-values from a series of statistical tests, requires independence of single test statistics in order to obtain asymptotic normality of the joint test statistic. The paper discusses the modification by Hartung (1999, Biometrical Journal, Vol. 41, pp. 849-855), which is designed to allow for a certain correlation matrix of the transformed P-values. First, the modified inverse normal method is shown here to be valid with more general correlation matrices. Secondly, a necessary and sufficient condition for (asymptotic) normality is provided, using the copula approach. Thirdly, applications to panels of cross-correlated time series, stationary as well as integrated, are considered. The behaviour of the modified inverse normal method is quantified by means of Monte Carlo experiments.
Oxford Bulletin of Economics and Statistics, 2014
ABSTRACT Ordinary fan charts consist of symmetric marginal forecast intervals, and do not take in... more ABSTRACT Ordinary fan charts consist of symmetric marginal forecast intervals, and do not take into consideration the concrete loss function of the user of the forecast. The note shows how to build fan charts that have exact joint coverage even under asymmetric loss, and maintain at the same time the intuition conveyed by ordinary fan charts. The proposed method is computationally simple, and easily implemented with any loss function. The differences between the information conveyed by fan charts with or without asymmetries, and with or without exact joint coverage, are illustrated with a Bayesian forecast exercise of US GDP growth rates.
Studying annual growth rates (seasonal difierences) in case of seasonal data produces much more p... more Studying annual growth rates (seasonal difierences) in case of seasonal data produces much more persistence, autocorrelation and stronger evidence in favour of a unit root than analyzing seasonal growth rates (ordinary difierences). First, this statement is quantifled theoretically. Second, it is supported experimentally with simulations, and, flnally, it is empirically illustrated with quarterly GDP de∞ators from 7 Eu- ropean economies.
Journal of Business & Economic Statistics, 2012
Uploads
Papers by Matei Demetrescu