                     Changes in version 0.9.14.9000                     

Update of version 0.9.13 that ...

                 Changes in version 0.9.13 (2026-09-14)                 

Update of version 0.9.12 that adds an estimation function to compute the
Anderson-Hsiao estimator for AR(1) panel data models, adjust the code of
functions NLIV, NLIV_t, and FDLS to allow for more general data
structures and unbalanced panels, and fixes typos in documentation.
Additionally, function NLIV.alt was renamed to NLIV_t.

NLIV

  - adjust code to allow for more general data structures
  - allow for unbalanced panels

NLIV_t

  - adjust code to allow for more general data structures
  - allow for unbalanced panels

FDLS

  - adjust code to allow for more general data structures
  - allow for unbalanced panels

AH81

  - closed form estimation functions for AR(1) panel data models
  - based on estimator proposed by Anderson and Hsiao (1981)

data.info

  - adjust code to allow for more general data structures

                 Changes in version 0.9.12 (2025-02-20)                 

Update of version 0.9.11 that includes some suggestions on code
improvements of Kevin Tappe (github user tappek) and corrects
computation of the onestep unadjusted standard errors and the serial
correlation test statistic. Adjustments to the instrument counts were
also made following pointers by Steve Bond. Additionally, the sargan.fct
was added which computes the sargan test statistic for the
overidentifying restrictions.

pdynmc

  - adjust computation of onestep unadjusted standard errors
  - add checks of input arguments
  - remove redundant parts of code
  - adjust function documentation and correct typos

mtest.fct

  - correct bugs in computation of the test statistic

wald.fct

  - remove redundant parts of code

sargan.fct

  - added to package for computing sargan test statistic

jtest.fct

  - warning message for non-spherical error terms added

plot.pdynmc

  - adjust function documentation
  - add further option for fire-plot

Z_i.fct

  - safeguard instrument count in internal helper function for the case
    when there are columns with zeros only

                 Changes in version 0.9.11 (2024-07-12)                 

Update of version 0.9.10 that adds three estimation functions for the
lag parameter of AR(1) panel data models. Additionally, the update
allows for user-specified dummy matrix in estimation function. For this
purpose, the internal helper function corSparse which was adopted from
package 'qlcMatrix' in a previous function update was adjusted.
Additionally, the argument checks of the estimation function were
updated and an option to collapse the moment conditions was added.

pdynmc

  - argument checks of estimation function updated
  - option for collapsing moment conditions added

NLIV

  - closed form estimation function for AR(1) panel data models
  - based on original version of Ahn and Schmidt (1995) moment
    conditions

NLIV.alt

  - closed form estimation functions for AR(1) panel data models
  - based on alternative formulation of Ahn and Schmidt (1995) moment
    conditions

FDLS

  - closed form estimation functions for AR(1) panel data models
  - based on estimator proposed by Han and Phillips (2010)

corSparse

  - internal helper function updated for checking for collinearities in
    dummy part of instrument matrix

                 Changes in version 0.9.10 (2023-11-24)                 

Update of version 0.9.9 that generalizes functionality of functions for
exploratory analysis of panel data. The function corSparse from package
'qlcMatrix' was added as internal helper function, as the aforementioned
package was scheduled to be moved from CRAN to the archive
by 2023-11-29. Additionally, bug fixes are provided for the estimation
function and the documentation of the package is adjusted according to
the new CRAN recommendation.

pdynmc

  - adjust check-related bug when instrumenting endogenous covariates

pDensTime.plot

  - ... argument added to function
  - adjust scaling of abscissa for general time periods
  - allow for user defined axis labels

corSparse

  - new internal helper function copied from package 'qlcMatrix', which
    is scheduled to be archived

                 Changes in version 0.9.9 (2023-06-06)                  

Update of version 0.9.8 which adds new function for visualization of
evolution of empirical density of a variable of interest over
longitudinal dimension of a panel dataset. Additionally, typos in
description of cigDemand dataset are adjusted and further information is
added to summary of `pdynmc' objects.

pdynmc

  - Summary function adjusted to add clarification on type of estimation

functions for exploratory analysis of panel data added

  - pDensTime.plot: : Visualizes empirical density of column of panel
    dataset

                 Changes in version 0.9.8 (2022-08-30)                  

Update of version 0.9.7 which adds functionality for excluding the
lagged dependent variable from the right-hand-side of the equation.
Additionally, the update adds the published version of the article as
vignette, ensures correct rendering of the package documentation (thanks
to Kevin Tappe), and corrects minor bugs in the estimation function
(thanks to Github user Dazhwu).

pdynmc

  - Add flexibility for excluding the lagged dependent variable from the
    right-hand-side of the model equation (by setting function argument
    lagTerms.y = 0).
  - Adjust documentation to reflect the new feature.
  - Adjust further controls part of instrument matrix when using
    instruments from further covariates (function arguments: include.x,
    varname.reg.pre, varname.reg.ex)
  - Adjust internal helper functions (Wonestep.fct, sub.clForm.fct)
    correspondingly

                 Changes in version 0.9.7 (2022-03-24)                  

Update of version 0.9.6 which updates the estimation function, the
functions for visualizing the panel data structure, and adds two
datasets to the package. The functionality for deriving instruments and
estimating parameters: Covariates for which no parameters are estimated,
but from which instruments are derived and covariates for which
parameters for which parameters are estimated, but from which no
instruments are derived.

pdynmc

  - Add flexibility for covariates for which no parameters are
    estimated, but which are used as instruments (function arguments:
    include.x.instr, varname.reg.instr)
  - Add flexibility for covariates for which parameters are estimated,
    but which are not used as instruments (function arguments:
    include.x.toInstr, varname.reg.toInstr)
  - Adjust code for further data structures
  - Update function documentation
  - Fix bug in function that creates instrument matrix

data.info

  - Adjust code for further data structures

struc.UPD.plot

  - Adjust code for further data structures

datasets

  - Add dataset of Arellano and Bond (1991) on employment by firms
    located in the UK
  - Add dataset of Stock and Watson (2003) on cigarette consumption in
    the US

                 Changes in version 0.9.6 (2021-10-14)                  

Minor update of version 0.9.5 which adds doubly corrected standard
errors. Also, commits and suggestions of github user tappek are added.
Additionally, the compatibility of the estimation function with further
input data structures is improved and a bug in the estimation function
when multiple instruments from non-lagged dependent endogenous
covariates are derived is corrected.

pdynmc

  - Implement doubly corrected standard errors
  - Adjust computation of unadjusted and corrected standard errors
  - Ensure compatibility with tibble data frames
  - Correct bug in deriving instruments from multiple endogenous
    covariates

wald.fct

  - Adjust structure of function according to S3 class htest.object

jtest.fct

  - Adjust structure of function according to S3 class htest.object
  - Minor adjustment to avoid partial argument matching

mtest.fct

  - Adjust structure of function according to S3 class htest.object
  - Set default for argument t.order
  - Minor adjustment to avoid partial argument matching

                 Changes in version 0.9.5 (2021-08-13)                  

Minor update of version 0.9.4 which adds further functionality and
argument checks to estimation function. Additionally, the computation
underlying non-robust two-step standard errors is adjusted (option
accessible by changing argument "std.err" from its default to "std.err =
unadjusted") and the functions for deriving instruments from further
exogenous covariates were adjusted to comply with data requirements.

pdynmc

  - Allow lagged dependent variable to be instrumented
  - Extend checks of function arguments
  - Fix bug in computation of "unadjusted" two-step standard errors
  - Ensure that instruments derived from further exogenous covariates
    comply with data requirements

                 Changes in version 0.9.4 (2021-06-14)                  

Minor update of version 0.9.3 in which package DESCRIPTION, CITATION,
and documentation is adjusted and two further package vignettes are
added.

new vignettes

  - pdynmc-intro
  - pdynmc_introLong
  - pdynmc-pres-in-a-nutshell

                 Changes in version 0.9.3 (2020-12-04)                  

Minor update of version 0.9.2 that fixes minor bugs in estimation
function and helper functions for setting up instrument and weighting
matrix; additionally, coefficient path plots are added to the plot
method and existing plot methods are adjusted.

pdynmc

  - Adjust default arguments and corresponding documentation
  - Fix bug in setting up of weighting matrix
  - Fix bug when imposing maximum lags from which instruments are
    derived
  - Add check functions
  - Adjust checks when setting up instrument matrix
  - Residuals and fitted values are added to pdynmc object (besides the
    previously available internal residuals and fitted values).

plot.pdynmc

  - Plotting coefficient paths across GMM iterations (including
    approximate standard errors).
  - Plot methods for creating fitted vs. residual plots and coefficient
    range plots are adjusted.

                 Changes in version 0.9.2 (2020-09-16)                  

Minor update of version 0.9.1 that adds additional functionality for
setting up the instrument matrix and adjusts the instrument count
displayed by the summary function.

pdynmc

  - Include additional option to limit expansion of the instrument set
    when deriving instrument from non-lagged dependent strictly
    exogenous covariates.

print.summary.pdynmc

  - Adjust displayed instrument count.

                 Changes in version 0.9.1 (2020-07-26)                  

Minor update of version 0.9.0 that fixes a bug in the estimation
function, adjusts matrix calculations to achieve minor speed
improvements, and robustifies general linear hypothesis testing
functionality.

pdynmc

  - fix bug that appeared when deriving moment conditions from the
    explanatory variables besides the lagged dependent variable (thanks
    to Massimo Giannini for pointing this out).
  - adjust matrix calculations to achieve minor speed improvements
  - adjust helper functions that allow limiting the lag range

wald.fct

  - Robustify wald.fct by using generalized inverse in inversion of
    covariance matrix.

                 Changes in version 0.9.0 (2020-05-07)                  

Update of version 0.8.0 that includes visualizations for fitted model
objects (coefficient-range plots for two-step and iterated estimation
and plots of fitted values vs. residuals) and panel data structure

functions for exploratory analysis of panel data added

  - data.info: Returns information on structure of a balanced/unbalanced
    panel dataset
  - strucUPD.plot: Visualizes structure of unbalanced panel data

generic functions added

  - ninst: Returns the number of instruments of a fitted model
  - optmIn: Returns input parameters used in numeric optimization of a
    fitted model
  - wmat: Returns weighting matrix of a fitted model

methods added

  - case.names.pdynmc: Returns variable names of cross-sectional and
    longitudinal identifiers of a fitted model
  - coef.pdynmc: Returns coefficient estimates of a fitted model
  - dummy.coef.pdynmc: Returns time dummy coefficient estimates of a
    fitted model
  - model.matrix.pdynmc: Returns instrument matrix of a fitted model
  - ninst.pdynmc: Returns instrument count of a fitted model
  - nobs.pdynmc: Returns number of cross-sectional and longitudinal
    observations of a fitted model
  - optimIn.pdynmc: Returns input parameters used in numeric
    optimization of a fitted model
  - plot.pdynmc: Plot methods for fitted model; returns plot of fitted
    values vs. residuals (default) or coefficient range across
    iterations of the estimation procedure (two-step or iterated
    estimation)
  - print.pdynmc: Print fitted model object in console
  - variable.names.pdynmc: Returns vector with names of explanatory
    variables of a fitted model
