We introduce CRE-GMM, a new estimator that exploits correlated random effects (CRE) within the generalised method of moments on level equations (GMMlev) in a dynamic (but also static) model on panel data. Unlike GMM-dif, it allows the estimation of the effects of measurable time-invariant covariates and, compared to GMM-sys, makes efficient use of all available information. CRE-GMM considers explanatory variables that may be affected by double endogeneity (correlation with individual heterogeneity and idiosyncratic shocks), models initial conditions and improves inference. Monte Carlo simulations validate CRE-GMM across panel types and endogeneity scenarios. Empirical applications to R&D, production, and wage functions illustrate the advantages of CRE-GMM.
Bontempi, M.E., Ditzen, J. (2025). The Correlated Random Effects GMM-Level Estimation: Monte Carlo Evidence and Empirical Applications. Springer Nature Switzerland AG 2025 : Springer Cham [10.1007/978-3-031-92699-0_11].
The Correlated Random Effects GMM-Level Estimation: Monte Carlo Evidence and Empirical Applications
Bontempi M. E.
;
2025
Abstract
We introduce CRE-GMM, a new estimator that exploits correlated random effects (CRE) within the generalised method of moments on level equations (GMMlev) in a dynamic (but also static) model on panel data. Unlike GMM-dif, it allows the estimation of the effects of measurable time-invariant covariates and, compared to GMM-sys, makes efficient use of all available information. CRE-GMM considers explanatory variables that may be affected by double endogeneity (correlation with individual heterogeneity and idiosyncratic shocks), models initial conditions and improves inference. Monte Carlo simulations validate CRE-GMM across panel types and endogeneity scenarios. Empirical applications to R&D, production, and wage functions illustrate the advantages of CRE-GMM.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
