Test of hypotheses in panel data models when the regressor and disturbances are possibly non-stationary

Under a Mundlak-type correlated random effect (CRE) specification, we first show that the average likelihood of a parametric nonlinear panel data model is the convolution of the conditional distribution of the model and the distribution of the unobserved heterogeneity. Hence, the distribution of the unobserved heterogeneity can be recovered by means of a Fourier transformation without imposing a distributional assumption on the CRE specification. We subsequently construct a semiparametric family of average likelihood functions of observables by combining the conditional distribution of the model and the recovered distribution of the unobserved heterogeneity, and show that the parameters in the nonlinear panel data model and in the CRE specification are identifiable. Based on the identification result, we propose a sieve maximum likelihood estimator. Compared with the conventional parametric CRE approaches, the advantage of our method is that it is not subject to misspecification on the distribution of the CRE. Furthermore, we show that the average partial effects are identifiable and extend our results to dynamic nonlinear panel data models.

Download Full-text

Spatiotemporal Econometrics Models for Old Age Mortality in Europe

Mathematics ◽

10.3390/math9091061 ◽

2021 ◽

Vol 9 (9) ◽

pp. 1061

Author(s):

Patricia Carracedo ◽

Ana Debón

Keyword(s):

Panel Data ◽

Goodness Of Fit ◽

Population Aging ◽

European Population ◽

Data Models ◽

European Countries ◽

Mortality Data ◽

Panel Data Models ◽

Estimation Of Parameters ◽

Advantages And Disadvantages

In the past decade, panel data models using time-series observations of several geographical units have become popular due to the availability of software able to implement them. The aim of this study is an updated comparison of estimation techniques between the implementations of spatiotemporal panel data models across MATLAB and R softwares in order to fit real mortality data. The case study used concerns the male and female mortality of the aged population of European countries. Mortality is quantified with the Comparative Mortality Figure, which is the most suitable statistic for comparing mortality by sex over space when detailed specific mortality is available for each studied population. The spatial dependence between the 26 European countries and their neighbors during 1995–2012 was confirmed through the Global Moran Index and the spatiotemporal panel data models. For this reason, it can be said that mortality in European population aging not only depends on differences in the health systems, which are subject to national discretion but also on supra-national developments. Finally, we conclude that although both programs seem similar, there are some differences in the estimation of parameters and goodness of fit measures being more reliable MATLAB. These differences have been justified by detailing the advantages and disadvantages of using each of them.

Download Full-text

Fitting partially linear functional-coefficient panel-data models with Stata

The Stata Journal Promoting communications on statistics and Stata ◽

10.1177/1536867x20976339 ◽

2020 ◽

Vol 20 (4) ◽

pp. 976-998

Author(s):

Kerui Du ◽

Yonghui Zhang ◽

Qiankun Zhou

Keyword(s):

Monte Carlo ◽

Panel Data ◽

Fixed Effects ◽

Linear Functional ◽

Semiparametric Estimation ◽

Data Models ◽

Panel Data Models ◽

Panel Models ◽

Partially Linear ◽

Functional Coefficient

In this article, we describe the implementation of fitting partially linear functional-coefficient panel models with fixed effects proposed by An, Hsiao, and Li [2016, Semiparametric estimation of partially linear varying coefficient panel data models in Essays in Honor of Aman Ullah ( Advances in Econometrics, Volume 36)] and Zhang and Zhou (Forthcoming, Econometric Reviews). Three new commands xtplfc, ivxtplfc, and xtdplfc are introduced and illustrated through Monte Carlo simulations to exemplify the effectiveness of these estimators.

Download Full-text

Backward mean transformation in unit root panel data models

Economics Letters ◽

10.1016/j.econlet.2021.109780 ◽

2021 ◽

Vol 201 ◽

pp. 109780

Author(s):

Artūras Juodis ◽

Rutger W. Poldermans

Keyword(s):

Panel Data ◽

Unit Root ◽

Data Models ◽

Panel Data Models

Download Full-text

Robust likelihood estimation of dynamic panel data models

Journal of Econometrics ◽

10.1016/j.jeconom.2021.03.005 ◽

2021 ◽

Author(s):

Javier Alvarez ◽

Manuel Arellano

Keyword(s):

Panel Data ◽

Likelihood Estimation ◽

Dynamic Panel Data ◽

Data Models ◽

Panel Data Models ◽

Dynamic Panel ◽

Dynamic Panel Data Models ◽

Robust Likelihood

Download Full-text

Estimating latent group structure in time-varying coefficient panel data models

Econometrics Journal ◽

10.1093/ectj/utz008 ◽

2019 ◽

Author(s):

Jia Chen

Keyword(s):

Panel Data ◽

Data Models ◽

Panel Data Models ◽

Time Varying ◽

Clustering Method ◽

Varying Coefficient ◽

Varying Coefficients ◽

Time Varying Coefficients ◽

Latent Group ◽

Group Structures

Summary This paper studies the estimation of latent group structures in heterogeneous time-varying coefficient panel data models. While allowing the coefficient functions to vary over cross-sections provides a good way to model cross-sectional heterogeneity, it reduces the degree of freedom and leads to poor estimation accuracy when the time-series length is short. On the other hand, in a lot of empirical studies, it is not uncommon to find that heterogeneous coefficients exhibit group structures where coefficients belonging to the same group are similar or identical. This paper aims to provide an easy and straightforward approach for estimating the underlying latent groups. This approach is based on the hierarchical agglomerative clustering (HAC) of kernel estimates of the heterogeneous time-varying coefficients when the number of groups is known. We establish the consistency of this clustering method and also propose a generalised information criterion for estimating the number of groups when it is unknown. Simulation studies are carried out to examine the finite-sample properties of the proposed clustering method as well as the post-clustering estimation of the group-specific time-varying coefficients. The simulation results show that our methods give comparable performance to the penalised-sieve-estimation-based classifier-LASSO approach by Su et al. (2018), but are computationally easier. An application to a panel study of economic growth is also provided.

Download Full-text