plot.pdynmc: Plot Coefficient Estimates and Corresponding Ranges of Fitted...

In pdynmc: Moment Condition Based Estimation of Linear Dynamic Panel Data Models

Plot Coefficient Estimates and Corresponding Ranges of Fitted Model.

Description

plot.pdynmc Plot methods for objects of class 'pdynmc'. The available plot options visualize: Fitted values versus residuals, coefficient ranges across GMM iterations, coefficient paths and objective function values across GMM iterations as proposed by \insertCiteHanLee2020inference;textualpdynmc.

Usage

## S3 method for class 'pdynmc'
plot(
  x,
  type = "fire",
  include.dum = FALSE,
  include.fur.con = FALSE,
  col.coefRange = 1,
  col.coefInitial = "darkgrey",
  col.coefEst = "royalblue",
  omit1step = FALSE,
  boxplot.coef = FALSE,
  co = NULL,
  add.se.approx = NULL,
  conf.lev = 0.95,
  ...
)

Arguments

x	An object of class 'pdynmc'. The function requires twostep or iterative GMM estimates.
type	Whether to plot fitted values against residuals (argument 'fire'; default), coefficient ranges (argument 'coef.range'; this requires twostep or iterative GMM estimates), path of coefficient estimates across GMM iterations (argument 'coef.path'; this requires twostep or iterative GMM estimates).
include.dum	Include estimates of parameters corresponding to time dummies (defaults to 'black'; requires 'type = coef.range').
include.fur.con	Include estimates of parameters corresponding to further controls (defaults to 'FALSE'; requires 'type = coef.range').
col.coefRange	Specify color for plotting range of coefficient estimates (defaults to 'NULL'; requires 'type = coef.range').
col.coefInitial	Specify color for plotting initial coefficient estimates (defaults to 'darkgrey'; requires 'type = coef.range').
col.coefEst	Specify color for plotting coefficient estimate (defaults to 'royalblue'; requires 'type = coef.range').
omit1step	Omit coefficient estimates from one-step GMM estimation in coefficient range plot. The argument can after obtaining coefficient estimates from numerical optimization methods to exclude the randomly drawn starting values from the plotted coefficient range (defaults to 'FALSE'). Set to 'TRUE' to exert the option; this argument requires iterative GMM estimates and argument 'type = coef.range'.
boxplot.coef	Whether to draw boxplots for coefficient estimates (defaults to 'FALSE'); requires iterative GMM with at least 10 iterations and argument 'type = coef.range'. Proceed with caution as this argument is experimental.
co	Character string denoting the variable name(s) for which to plot the path of coefficient estimate(s) across GMM iterations (defaults to 'NULL') as proposed in \insertCiteHanLee2020inference;textualpdynmc; if no coefficient name is given, all coefficient paths are plotted; requires at least two iterations and argument 'type = coef.path'.
add.se.approx	A logical variable indicating if standard errors should be added to the plot of the path of coefficient estimate(s) across GMM iterations (defaults to 'NULL'); requires at least two iterations and argument 'type = coef.path'. This option is only available when plotting a single coefficient path (i.e., when 'co' contains only a single variable name).
conf.lev	A numeric variable indicating the confidence level for approximating standard errors in the plot of the path of coefficient estimate(s) across GMM iterations (defaults to 0.95; sensible values lie in the interval ]0,1[); requires argument 'type = coef.path' and argument 'add.se.approx = TRUE'.
...	further arguments.

Value

Plot fitted values against residuals ('type = fire') or coefficient estimates and coefficient estimate ranges ('type = coef.range') for object of class 'pdynmc'. The latter plot requires twostep or iterative GMM estimates.

Author(s)

Markus Fritsch and Joachim Schnurbus

References

\insertAllCited

See Also

pdynmc for fitting a linear dynamic panel data model.

Examples

## Load data
data(ABdata, package = "pdynmc")
dat <- ABdata
dat[,c(4:7)] <- log(dat[,c(4:7)])
dat <- dat[c(1:140), ]

## Code example
m1 <- pdynmc(dat = dat, varname.i = "firm", varname.t = "year",
    use.mc.diff = TRUE, use.mc.lev = FALSE, use.mc.nonlin = FALSE,
    include.y = TRUE, varname.y = "emp", lagTerms.y = 2,
    fur.con = TRUE, fur.con.diff = TRUE, fur.con.lev = FALSE,
    varname.reg.fur = c("wage", "capital", "output"), lagTerms.reg.fur = c(1,2,2),
    include.dum = TRUE, dum.diff = TRUE, dum.lev = FALSE, varname.dum = "year",
    w.mat = "iid.err", std.err = "corrected", estimation = "twostep",
    opt.meth = "none")
plot(m1)
plot(m1, type = "coef.range")
plot(m1, type = "coef.path")


## Load data
 data(ABdata, package = "pdynmc")
 dat <- ABdata
 dat[,c(4:7)] <- log(dat[,c(4:7)])

## Further code example
 m1 <- pdynmc(dat = dat, varname.i = "firm", varname.t = "year",
    use.mc.diff = TRUE, use.mc.lev = FALSE, use.mc.nonlin = FALSE,
    include.y = TRUE, varname.y = "emp", lagTerms.y = 2,
    fur.con = TRUE, fur.con.diff = TRUE, fur.con.lev = FALSE,
    varname.reg.fur = c("wage", "capital", "output"), lagTerms.reg.fur = c(1,2,2),
    include.dum = TRUE, dum.diff = TRUE, dum.lev = FALSE, varname.dum = "year",
    w.mat = "iid.err", std.err = "corrected", estimation = "iterative",
    opt.meth = "none")
 plot(m1)
 plot(m1, type = "coef.range")
 plot(m1, type = "coef.path")

pdynmc documentation built on Sept. 12, 2024, 7:42 a.m.