J. Paul Elhorst

Specification and estimation of spatial panel data models

This article provides a survey of the specification and estimation of spatial panel data models. These models include spatial error autocorrelation, or the specification is extended with a spatially lagged dependent variable. In particular, the author focuses on the specification and estimation of four panel data models commonly used in applied research: the fixed effects model, the random effects model, the fixed coefficients model, and the random coefficients model. The survey discusses the asymptotic properties of the estimators and provides guidance with respect to the estimation procedures, which should be useful for practitioners.

Spatial Panel Data Models

In recent years, the spatial econometrics literature has exhibited a growing interest in the specification and estimation of econometric relationships based on spatial panels. Spatial panels typically refer to data containing time series observations of a number of spatial units (zip codes, municipalities, regions, states, jurisdictions, countries, etc.). This interest can be explained by the fact that panel data offer researchers extended modeling possibilities as compared to the single equation cross-sectional setting, which was the primary focus of the spatial econometrics literature for a long time. Panel data are generally more informative, and they contain more variation and less collinearity among the variables. The use of panel data results in a greater availability of degrees of freedom, and hence increases efficiency in the estimation. Panel data also allow for the specification of more complicated behavioral hypotheses, including effects that cannot be addressed using pure cross-sectional data (see Hsiao 2005 for more details).

Matlab Software for Spatial Panels

Elhorst provides Matlab routines to estimate spatial panel data models at his website. This article extends these routines to include the bias correction procedure proposed by Lee and Yu if the spatial panel data model contains spatial and/or time-period fixed effects, the direct and indirect effects estimates of the explanatory variables proposed by LeSage and Pace, and a selection framework to determine which spatial panel data model best describes the data. To demonstrate these routines in an empirical setting, a demand model for cigarettes is estimated based on panel data from forty-six US states over the period 1963–1992.

Dynamic spatial panels: models, methods, and inferences

This paper provides a survey of the existing literature on the specification and estimation of dynamic spatial panel data models, a collection of models for spatial panels extended to include one or more of the following variables and/or error terms: a dependent variable lagged in time, a dependent variable lagged in space, a dependent variable lagged in both space and time, independent variables lagged in time, independent variables lagged in space, serial error autocorrelation, spatial error autocorrelation, spatial-specific and time-period-specific effects. The survey also examines the reasoning behind different model specifications and the purposes for which they can be used, which should be useful for practitioners.

Evidence of political yardstick competition in France using a two-regime spatial Durbin model with fixed effects

This research proposes a two-regime spatial Durbin model with spatial and time-period fixed effects to test for political yardstick competition and exclude any other explanation that might produce spatial interaction effects among the dependent variable, the independent variables, or the error term. The study also derives the maximum likelihood estimator and variance-covariance matrix of the parameters of this model. Data pertaining to welfare spending by 93 departments in France during 1992-2000 provide significant empirical evidence in support of political yardstick competition. Departments governed by a small political majority mimic neighboring expenditures on welfare to a greater extent than do departments governed by a large political majority. Copyright (c) 2009, Wiley Periodicals, Inc.

NEW EVIDENCE ON THE WAGE CURVE

Following Blanchflower and Oswald, a “wage curve” describes the wage level as a downward-sloping convex curve of the regional unemployment rate. This article makes two major contributions in the analysis of the wage curve. First, it is recognized that potential endogeneity of the regional unemployment rate should be subject to testing not only in combination with regional-specific effects but also in combination with time-specific effects. For this purpose, the authors develop a new estimator, the spatial first difference 2SLS estimator. Second, it is recognized that wages may not only respond to the regional but also to the national unemployment rate. In the empirical analysis, the wage curve for East Germany is estimated using a comprehensive database that provides panel data classified into 114 administrative districts during the 1993 to 1999 period.

A Simultaneous Equations Model of Fiscal Policy Interactions

Existing studies of fiscal policy interactions are based on single equation (SE) models of either taxation or expenditures, without specifying the underlying social welfare function, without taking account of budget constraints and without allowing for cost differences between jurisdictions. Taking all this into account, we derive an extended version of the linear expenditure system with policy interaction effects. We use this system to simultaneously estimate interactions in both taxation and different spending categories among Dutch municipalities. Our interaction parameters tend to be higher than those estimated using conventional SE models.

Forecasting the impact of transport improvements on commuting and residential choice

This paper develops a probabilistic, competing-destinations, assignment model that predicts changes in the spatial pattern of the working population as a result of transport improvements. The choice of residence is explained by a new non-parametric model, which represents an alternative to the popular multinominal logit model. Travel times between zones are approximated by a normal distribution function with different mean and variance for each pair of zones, whereas previous models only use average travel times. The model’s forecast error of the spatial distribution of the Dutch working population is 7% when tested on 1998 base-year data. To incorporate endogenous changes in its causal variables, an almost ideal demand system is estimated to explain the choice of transport mode, and a new economic geography inter-industry model (RAEM) is estimated to explain the spatial distribution of employment. In the application, the model is used to forecast the impact of six mutually exclusive Dutch core-periphery railway proposals in the projection year 2020.

A spatiotemporal analysis of aggregate labour force behaviour by sex and age across the European Union

This study investigates the causes of variation in age-specific male and female labour force participation rates using annual data from 154 regions across ten European Union member states for the period 1983–1997. Regional participation rates appear to be strongly correlated in time, weakly correlated in space and to parallel their national counterparts. An econometric model is designed consistent with these empirical findings. To control for potential endogeneity of the explanatory variables, we use an instrumental variables estimation scheme based on a matrix exponential spatial specification of the error terms. Many empirical studies of aggregate labour force behaviour have ignored population distribution effects, relying instead on the representative-agent paradigm. In order for representative-agent models to accurately describe aggregate behaviour, all marginal reactions of individuals to changes in aggregate variables must be identical. It turns out that this condition cannot apply to individuals across different sex/age groups.

Urban and rural population growth in a spatial panel of municipalities

ABSTRACT Urban and rural population growth in a spatial panel of municipalities. Regional Studies. Using Bayesian posterior model probabilities and data pertaining to 3659 Brazilian minimum comparable areas (MCAs) over the period 1970–2010, two theoretical settings of population growth dynamics resulting in two spatial econometric specifications in combination with a wide range of potential neighbourhood matrices are tested against each other. The best performing combination counts five determinants producing significant long-term spatial spillover effects. Ignoring these spillovers, as many previous population growth studies have done, is shown to underestimate their impact and thus the effectiveness of policy measures acting on these determinants.