fbardl: Fourier Bootstrap ARDL Cointegration Test

R implementation of the Fourier Bootstrap ARDL (FBARDL) bounds testing approach for cointegration analysis.

Overview

The fbardl package combines: - Pesaran, Shin & Smith (2001) ARDL bounds testing framework - Fourier terms to capture smooth structural breaks (Yilanci et al., 2020) - Bootstrap critical values for robust inference (McNown et al., 2018; Bertelli et al., 2020)

Installation

# Install from source
install.packages("fbardl_1.0.0.tar.gz", repos = NULL, type = "source")

# Or using devtools
devtools::install_github("muhammedalkhalaf/fbardl")

Usage

library(fbardl)

# Load example data
data(fbardl_data)

# Basic Fourier ARDL test
result <- fbardl(y ~ x1 + x2, data = fbardl_data, type = "fardl")
summary(result)

# Bootstrap ARDL (McNown approach)
result_boot <- fbardl(y ~ x1 + x2, data = fbardl_data,
                      type = "fbardl_mcnown", reps = 999)
summary(result_boot)

# Without Fourier terms
result_nofourier <- fbardl(y ~ x1 + x2, data = fbardl_data, fourier = FALSE)

Test Types

Type Description
"fardl" Fourier ARDL with PSS bounds test (default)
"fbardl_mcnown" Bootstrap ARDL (McNown, Sam & Goh, 2018)
"fbardl_bvz" Bootstrap ARDL (Bertelli, Vacca & Zoia, 2020)

Features

Output

The function returns an object of class "fbardl" containing: - Model coefficients and standard errors - Long-run and short-run coefficient estimates - Cointegration test statistics and p-values - Diagnostic test results - Model fit statistics (R², AIC, BIC)

References

Author

Dr. Merwan Roudane (merwanroudane920@gmail.com)

License

GPL-3