XTPVARCOINT: Stata module to perform Panel VAR Modeling with Cointegration, Structural Breaks, and Cross-Sectional Dependence

xtpvarcoint implements a comprehensive econometric toolkit for vector autoregressive (VAR) modeling in heterogeneous panels. It is a Stata port of the R package pvars and provides: Panel cointegration rank tests based on the Johansen procedure (Larsson et al. 2001), the Breitung (2005) two-step estimator, the Saikkonen-Luetkepohl procedure (Arsova and Oersal 2018), and the correlation-augmented inverse normal test robust to cross-sectional dependence (Arsova and Oersal 2021). Panel VAR/VECM estimation using the mean-group (MG) and pooled mean-group (PMG) estimators with optional Kilian (1998) bias correction and factor-augmented specifications for cross-sectional dependence. Panel SVAR identification via Cholesky decomposition, Blanchard-Quah long-run restrictions, proxy/IV variables, distance covariance (DC), and Cramer-von Mises (CVM) independence criteria. Specification tools for determining the number of common factors (Onatski 2010, Ahn and Horenstein 2013, Bai and Ng 2002) and the optimal lag order based on AIC, HQC, SIC, and FPE criteria. Impulse response functions, {bf:forecast error variance decomposition}, and bootstrap confidence intervals via the panel moving-block bootstrap. Publication-quality plots for IRFs with confidence bands, FEVD stacked area charts, and companion matrix eigenvalue stability diagrams. The data must be xtset before calling xtpvarcoint. Individual-specific lag orders are supported by specifying a numlist in lags().