Salvato in:
Dettagli Bibliografici
Autori principali: Dey, Subhodeep, Basak, Gopal K., Das, Samarjit
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:https://arxiv.org/abs/2408.13514
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
Sommario:
  • We re-investigate the asymptotic properties of the traditional OLS (pooled) estimator, $\hatβ _P$, in the context of cluster dependence. The present study considers various scenarios under various restrictions on the cluster sizes and number of clusters. It is shown that $\hatβ_P$ could be inconsistent in many realistic situations. We propose a simple estimator, $\hatβ_A$ based on data averaging. The asymptotic properties of $\hatβ_A$ are studied. It is shown that $\hatβ_A$ is consistent even when $\hatβ_P$ is inconsistent. It is further shown that the proposed estimator $\hatβ_A$ is more efficient than $\hatβ_P$ in many practical scenarios. As a consequence of averaging, we show that $\hatβ_A$ retains consistency, asymptotic normality under classical measurement error problem circumventing the use of Instrumental Variables (IV). A detailed simulation study shows the efficacy of $\hatβ_A$. It is also seen that $\hatβ_A$ yields better goodness of fit.