Modified Instruments: An Iterative Process for Reducing Endogeneity Bias Using Large Samples

Nicolás Ronderos Pulido ()

No 20049, Documentos de Economía from Universidad Javeriana - Bogotá

Abstract: This Article proposes an iterative process to reduce bias of the instrumental variable estimator using large samples. One can achieve bias reduction by modifying the data of an instrument. Success in reducing bias depends on the magnitude of exogeneity in the instrument. The common exogeneity between the instrument and the problem variable must be greater than the common endogeneity between the same variables. If the instrument is weakly exogenous, the iterative process will not affect the estimates. Empirically, the iterative process requires searching for parameters that allow minimizing bias. This paper presents a convergence algorithm to search for parameters. The search algorithm and statistical inference are based on bootstrapping techniques. The results are presented in a simulation context.

Keywords: instrumental variables; bias correction (search for similar items in EconPapers)
JEL-codes: C26 C40 (search for similar items in EconPapers)
Pages: 46
Date: 2022-04-27
New Economics Papers: this item is included in nep-ecm and nep-ore
References: Add references at CitEc
Citations:

Downloads: (external link)
https://drive.google.com/file/d/1wbSl0lswcNH58KMJJjZpVrtgrpbuA0zM/view

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:col:000108:020049

Access Statistics for this paper

More papers in Documentos de Economía from Universidad Javeriana - Bogotá Contact information at EDIRC.
Bibliographic data for series maintained by Mayerly Galindo Rodriguez ().