penfmbr: penfmbr: Plain FamaMacBeth and Penalized FamaMacBeth...

Description Details


For the penalized FamaMacBeth, the level tuning parameter is set to the average of the residual standard deviation from the first stage. The dimensionalty paramter is 4 by default, and individual penalty weights are calculated as the Fisher transform of partial correlations.


'returns' and 'factors' are two containers with equal number of observations (aligned) that can be coerced to a matrix. Both estimators return a crudely hacked together sublass of 'lm' that plays nicely with 'stargazer'.

'vcov' can return either Shanken or GMM standard errors. For a 'penfmb' instance, the bootstrapped shrinkage rate is returned.

nuffe/penfmbr documentation built on June 6, 2019, 11:06 p.m.