mle.tar

The MLE and EMM packages are similar, for many problems either
would work.  Both are misnomers because they cover a much
broader class of estimators than their names suggest.  The
difference between them is that EMM presumes that a model can be
simulated and MLE presumes that it cannot.

The main practical difference is that the heteroskedasticity
autoregressive consistent (HAC) estimator of the information
matrix that goes in the middle of a sandwich variance estimator
can be computed by parametric bootstrap if a model can be
simulated.  A sandwich variance has the inverse of the Hessian
of the log likelihood on either side and has a HAC estimate of
the variance of the gradient of the log likelihood in the
middle.  In both EMM and MLE the inverse of the Hessian is
estimated by the variance matrix of an MCMC chain.  EMM can
compute a bootstrap HAC from a simulation or a stationary
bootstrap.  MLE cannot and instead relies on user supplied
scores.

The design of the EMM code, although facilitating bootstrap
estimation of the HAC in the middle, makes it impossible to use
EMM for the Metropolis within Gibbs estimator for generalized
method of moments (GMM) estimation of models with latent
variables proposed by Gallant, Giacomini, and Ragusa (2013).
The technical reason is that EMM uses two instantiations of the
user supplied model (one for the MCMC chain and another for the
HAC in the middle) that would need to communicate with each
other to implement Metropolis within Gibbs but cannot within EMM
without offense to the principles of object oriented programming
and parallel-safe design.  MLE was written to fix this problem.

The package is written in C++.  It includes a parallelized
version that uses MPI, a user's guide, examples, and necessary
libraries.

This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License as
published by the Free Software Foundation; either version 2 of
the License, or (at your option) any later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public
License along with this program; if not, write to the Free
Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
Boston, MA 02110-1301 USA.