In franzmohr/prais: Prais-Winsten Estimator for AR(1) Serial Correlation

prais

Overview

prais implements the Prais-Winsten estimator for models with strictly exogenous regressors and AR(1) serial correlation of the errors.

Installation

CRAN

install.packages("prais")

Development version

# install.packages("devtools")
devtools::install_github("franzmohr/prais")

Usage

# Load the package
library(prais)

# Load the data
data("barium")

pw <- prais_winsten(lchnimp ~ lchempi + lgas + lrtwex + befile6 + affile6 + afdec6,
                    data = barium, index = "t")
summary(pw)

Robust standard errors

White's estimator

library(lmtest)

coeftest(pw, vcov. = vcovHC(pw, "HC1"))

Panel-corrected standard errors (PCSE)

Estimate a panel model, for which PCSE should be obtained.

# Example 2 in the documentation of Stata function xtpcse

# Load data
data <- haven::read_dta("http://www.stata-press.com/data/r14/grunfeld.dta")

# Estimate
x <- prais_winsten(invest ~ mvalue + kstock, data = data, index = c("company", "year"),
                   twostep = TRUE, panelwise = TRUE, rhoweight = "T1")

# Results
summary(x)

Obtain PCSE by using only those residuals from periods that are common to all panels by setting pairwise = FALSE.

coeftest(x, vcov. = vcovPC(x, pairwise = FALSE))

Obtain PCSE by using all observations that can be matched by period between two panels by setting pairwise = TRUE.

coeftest(x, vcov. = vcovPC(x, pairwise = TRUE))

References

Beck, N. L. and Katz, J. N. (1995): What to do (and not to do) with time-series cross-section data. American Political Science Review 89, 634-647.

Prais, S. J. and Winsten, C. B. (1954): Trend Estimators and Serial Correlation. Cowles Commission Discussion Paper, 383 (Chicago).

Wooldridge, J. M. (2016). Introductory Econometrics. A Modern Approach. 6th ed. Mason, OH: South-Western Cengage Learning.

franzmohr/prais documentation built on Nov. 26, 2024, 3:45 a.m.