The Bias of the Modified Limited Information Maximum Likelihood Estimator (MLIML) in Static Simultaneous Equation Models

Gareth Liu-Evans and Garry DA Phillips

No 202303, Working Papers from University of Liverpool, Department of Economics

Abstract: A higher-order approximation is made to the bias of the modified LIML (MLIML) estimator due to Fuller. It is demonstrated via simulation that the asymptotic approximation can be used to reduce estimation bias, including in cases where instrument strength is relatively weak, and that the approximation also mirrors the behaviour of the true bias. It is possible to see via the asymptotic approximation why MLIML estimation bias is often found to be very small in two equation models where the order of overidentification is small, and to predict, in simple models where the approximation is specialised, how the order of overidentification will relate nonlinearly to the bias. An asymptotic approximation is also obtained for the pseudo-bias of the LIML estimator. Finally, the bias-corrected MLIML estimator is used to re-examine the effect on the US college graduate wage premium of shifts in the relative supply of young college workers, following Fortin (2006).

Keywords: LIML; Modified LIML; 2SLS; bias approximation; bias correction (search for similar items in EconPapers)
Pages: 23 pages
Date: 2023
New Economics Papers: this item is included in nep-ecm
References: View references in EconPapers View complete reference list from CitEc
Citations:

Forthcoming

Downloads: (external link)
https://www.liverpool.ac.uk/media/livacuk/schoolof ... s/ECON,WP,202303.pdf First version, 2023 (application/pdf)

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:liv:livedp:202303

Access Statistics for this paper

More papers in Working Papers from University of Liverpool, Department of Economics Contact information at EDIRC.
Bibliographic data for series maintained by Rachel Slater ().