Entropy balancing as an estimation command

Ben Jann

No 39, University of Bern Social Sciences Working Papers from University of Bern, Department of Social Sciences

Abstract: Entropy balancing is a popular reweighting technique that provides an alternative to approaches such as, for example, inverse probability weighting (IPW) based on a logit or probit model. Even if the balancing weights resulting from the procedure will be of primary interest in most applications, it is noteworthy that entropy balancing can be represented as a simple regression-like model. An advantage of treating entropy balancing as a parametric model is that it clarifies how the reweighting affects statistical inference. In this article I present a new Stata command called -ebalfit- that estimates such a model including the variance-covariance matrix of the estimated coefficients. The balancing weights are then obtained as model predictions. Variance estimation is based on influence functions, which can be stored for further use, for example, to obtain consistent standard errors for statistics computed from the reweighted data.

Keywords: entropy balancing; reweighting; inverse probability weighting; IPW; influence function (search for similar items in EconPapers)
JEL-codes: C01 C21 C87 (search for similar items in EconPapers)
Pages: 31 pages
Date: 2021-08-03, Revised 2021-08-16
New Economics Papers: this item is included in nep-isf
References: View references in EconPapers View complete reference list from CitEc
Citations: View citations in EconPapers (3)

Downloads: (external link)
https://boris.unibe.ch/157883/15/jann-2021-ebalfit.pdf Revised version (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:bss:wpaper:39

Access Statistics for this paper

More papers in University of Bern Social Sciences Working Papers from University of Bern, Department of Social Sciences
Bibliographic data for series maintained by Ben Jann ().