XTCSPQARDL: Stata module to perform Cross-Sectionally Augmented Panel Quantile ARDL, Quantile CCE Mean Group, and Quantile CCE Pooled Mean Group Estimation

xtcspqardl implements two estimators for dynamic heterogeneous panel data models with cross-sectional dependence at the quantile level: CS-PQARDL (Cross-Sectionally Augmented Panel Quantile ARDL): Extends the panel quantile ARDL framework by augmenting unit-specific regressions with cross-sectional averages (CSA) of the dependent and independent variables, following Pesaran (2006) and Chudik and Pesaran (2015). This absorbs unobserved common factors that induce cross-sectional dependence, producing consistent estimates under factor structures. CSA enter the long-run equation (or ECT in ECM form), but are excluded from short-run dynamics. QCCEMG (Quantile CCE Mean Group): Implements the Quantile Common Correlated Effects Mean Group estimator of Harding, Lamarche, and Pesaran (2018). This is a dynamic panel quantile regression model where unobserved common factors are proxied by CSA and their lags. Unit-specific quantile regressions are averaged using the Mean Group approach, with inference based on the nonparametric between-group variance (Theorem 4 of HLP 2018). Both estimators handle heterogeneous slope coefficients across panels, permit quantile-varying parameters, and are robust to cross-sectional dependence generated by unobserved common factors. QCCEPMG (Quantile CCE Pooled Mean Group): Implements the pooled version of the QCCEMG estimator, following Pesaran (2006) CCEP methodology adapted to the quantile setting. Individual quantile regressions are estimated identically to QCCEMG, but the Mean Group averaging and inference follow the Pooled Mean Group approach. Under slope homogeneity, the QCCEPMG estimator achieves √(NT) convergence compared to √N for QCCEMG. All three estimators handle heterogeneous slope coefficients across panels, permit quantile-varying parameters, and are robust to cross-sectional dependence generated by unobserved common factors.