tsgarch

it does not (yet) implement all GARCH models in rugarch. FIGARCH, Multiplicative Component GARCH and Realized GARCH are not currently implemented.
it does not implement joint ARFIMA-GARCH estimation, but does allow for a jointly estimated ARMA(p,q)-X mean equation (via the arma and xreg arguments to garch_modelspec) in addition to the constant (since version 1.0.5). With so many options for modelling the conditional mean beyond low order ARMA dynamics, many of which are available in the tsmodels framework, it was decided to keep this package’s mean equation support deliberately limited rather than pursue full joint estimation of arbitrary mean dynamics with the variance equation. For a fully general mean equation, the 2 step estimation approach, whereby the residuals of the conditional mean are passed to the variance dynamics estimation, remains available and may be more flexible in what can be achieved, even if potentially less efficient for small sized datasets. Additionally, the ARCH-in-mean model is no longer available, as it was found to have very limited value within the tsmodels framework or in this author’s experience. A separate tsarma package for ARMA(p,q)-X models is however available.
it makes use of automatic differentiation (autodiff) during estimation, via the TMB package. This is in line with similar approaches in other models written in the tsmodels framework. Using autodiff allows for more confident estimation and more accurate standard errors.
it fully implements and correctly documents a number of sandwich estimators making use of the sandwich package framework (with methods for bread and estfun and meat/meat_HAC functions).
it makes use of S3 methods and classes, abandoning the S4 approach of rugarch. Additionally, while making use of standard methods from the stats package, some of the methods are based on those exported from tsmethods, consistent with other packages in the tsmodels framework.

Installation

The package can be installed from CRAN or the tsmodels github repo:

remotes::install_github("tsmodels/tsgarch", dependencies = TRUE)

Note, that in order to make use of symbolic output, flextable requires equatags to be installed which has a dependency on xlst which in turn has SystemRequirements libxslt. Therefore, if you are seeing NA printed in place of symbols, then it is likely that xlst is not installed.

Performance

Due to the heavy use of the data.table package for parameter tracking and indexing, I have seen significant speed improvements by setting the following :

options(datatable.optimize = 2L)