I recently used a R package CovTools in R with the command CovEst.2003LW(X), where X is your sample covariance matrix as an input, to compute the shrunk covariance matrix (an estimate that is closest to the true covariance matrix). For reference, the package resides here:


Can any portfolio optimization or R experts provide their insights into whether this package is trustable? The package is based on the famous paper "Honey, I Shrunk the Sample Covariance Matrix" by Ledoit and Wolf (2003).

  • 2
    $\begingroup$ If you do not get an answer here, but might want to ask at R-SIG-Finance (stat.ethz.ch/mailman/listinfo/r-sig-finance). Several packages on CRAN implement Ledoit--Wolf (e.g. RiskPortfolios::covEstimation or BurStFin::var.shrink.eqcor), so you also compare results. $\endgroup$ Nov 2, 2023 at 11:17

1 Answer 1


You can use the code provided from the authors directly. Michael Wolf has a whole library of examples here in different programming languages. For the "Honey, I Shrunk the Covariance Matrix" paper you can check out this Github repo.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.