System Estimation of Panel Data Models Under Long-Range Dependence

Abstract

Projections are carried out based onwhere bold indicates the vector of parameters with the critical parameter values being d max

Estimation of fractionally integrated panels with fixed effects and cross-section dependence

“…Our methodology for memory estimation consists in CSS estimation on the first differences of defactored variables, where projections are carried out on the sample means of differenced data, and slope estimation is carried out in the least-squares sense. While our methodology offers a general treatment for stationary and nonstationary indicators and works well in practice as indicated by Monte Carlo experiments, it can nevertheless be extended in the following directions: (a) Different estimation techniques, such as fixed effects and GMM, can be used under our setup as in Robinson and Velasco (2015); (b) The idiosyncratic shocks may be allowed to feature spatial dependence providing further insights in empirical analyses; (c) The independence assumption between the idiosyncratic shocks in the general model can be relaxed to allow for nonfactor endogeneity thereby leading to a cointegrated system analysis in the classical sense as in Ergemen (2015) who considers a less flexible modelization due to the lack of allowance of multiple covariates; (d) Panel unit-root and related hypothesis testing can be readily performed using our methodology, but it could also be interesting to develop tests that can detect breaks in the general model parameters; (e) Homogeneity tests on the slope parameters could be developed by comparing our mean group estimates with pooled estimates derived from a homogeneous version of our model.…”

Fundamental Responsiveness in European Electricity Prices

“…Panel data models can vary the pricing relationship between hourly trading sessions (this is the so-called slope heterogeneity property) by using regressors that are pertinent to each session (hour-specific regressors). Recent developments in panel data techniques, such as the common correlated effects estimator of Pesaran [30] and its refinements proposed by Ergemen [31] and Thomaidis and Biskas [29], make it possible to consistently estimate price responsiveness to fundamental variates, taking also into account unobserved crosscorrelation, long-range dependance and short-memory dynamics in innovations. Although the focus of interest may be the derivation of empirical fundamental pricing laws, all "secondary" aspects of the system dynamics can adversely affect the above task, often leading to inconsistent estimators for the slope coefficients.…”

Persistence Heterogeneity Testing in Panels with Interactive Fixed Effects

“…} , where f t could be the major source of persistence in data. The model could be complemented with the presence of incidental trends and other exogenous or endogenous observable regressor series, see Ergemen and Velasco (2017) and Ergemen (2017).…”

“…There is a recent literature on large N, T panel data models with long-range dependence constituting an alternative to AR specifications, see for example Robinson and Velasco (2015), Ergemen and Velasco (2017) and Ergemen (2017). Like models with AR dynamics, these models nest I(1) behaviour, but smoothly and thus the estimates of long-range dependence parameters are asymptotically normal unlike the non-standard asymptotics under non-stationary AR specifications.…”

“…Although classical AR panel literature uses both homogeneous and heterogeneous AR parameter specifications in modelling persistence, little or no justification is provided for these choices. In fractional panel data literature, Ergemen and Velasco (2017) and Ergemen (2017) allow memory parameters under their setups to vary across cross-section units arguing that this allows for a greater flexibility in suiting the presentation of data as well as describing the dynamics of different units more accurately rather than restricting all units to have the same √ NT convergence rate. We nevertheless discuss bias correction methods that relax the restriction that T should grow substantially faster than N in the joint asymptotics, which would not affect the estimation of the heterogeneous model.…”