Pooled Mean Group Estimation of Dynamic Heterogeneous Panels

Pooled Mean Group Estimation of Dynamic Heterogeneous Panels

November 1998 | Yongcheol Shin, Mohammad Hashem Pesaran, Ron P Smith
This paper proposes a new estimator for dynamic heterogeneous panels, called the Pooled Mean Group (PMG) estimator, which combines elements of both the Mean Group (MG) and pooled estimators. The PMG estimator constraints the long-run coefficients to be identical across groups but allows the short-run coefficients and error variances to differ. The authors derive the asymptotic distribution of the PMG estimators for both stationary and non-stationary regressors and provide empirical applications to aggregate consumption functions and energy demand functions. The PMG estimator is shown to be consistent and asymptotically normal, with smaller standard errors for the long-run coefficients compared to the MG estimator. The paper also discusses the choice between fixed and random effects models, the identification conditions, and the handling of heteroskedasticity and serial correlation.This paper proposes a new estimator for dynamic heterogeneous panels, called the Pooled Mean Group (PMG) estimator, which combines elements of both the Mean Group (MG) and pooled estimators. The PMG estimator constraints the long-run coefficients to be identical across groups but allows the short-run coefficients and error variances to differ. The authors derive the asymptotic distribution of the PMG estimators for both stationary and non-stationary regressors and provide empirical applications to aggregate consumption functions and energy demand functions. The PMG estimator is shown to be consistent and asymptotically normal, with smaller standard errors for the long-run coefficients compared to the MG estimator. The paper also discusses the choice between fixed and random effects models, the identification conditions, and the handling of heteroskedasticity and serial correlation.
Reach us at info@study.space
[slides and audio] Pooled Mean Group Estimation of Dynamic Heterogeneous Panels