Shrinking the covariance matrix

David J. Disatnik*, Simon Benninga

*Corresponding author for this work

Research output: Contribution to journalReview articlepeer-review

38 Scopus citations


The subject here is construction of the covariance matrix for portfolio optimization. In terms of the ex post standard deviation of the global minimum-variance portfolio, there is no statistically significant gain in using more sophisticated shrinkage estimators rather than simpler portfolios of estimators. This finding holds whether or not the investor imposes short sale constraints to prevent portfolio weights from being negative.

Original languageEnglish
Pages (from-to)55-63+6
JournalJournal of Portfolio Management
Issue number4
StatePublished - 2007


Dive into the research topics of 'Shrinking the covariance matrix'. Together they form a unique fingerprint.

Cite this