tests/testthat/test-lp_lin_iv.R
In AdaemmerP/lpirfs: Local Projections Impulse Response Functions
# Load package data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# 'Instrument' variable
instrument <- as.data.frame(ag_data$Gov[sample_start:sample_end])
test_that("Test whether data is a data.frame", {
endog_data <- as.matrix(endog_data)
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
'The data has to be a data.frame().', fixed = TRUE)
})
test_that("Test whether instrument is given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = NULL,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
'You have to provide an instrument to shock with.', fixed = TRUE)
})
test_that("Test whether lags_endog_lin is falsely NaN", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = NaN,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
'"lags_endog_lin" can only be NaN if a lag length criterion is given.', fixed = TRUE)
})
test_that("Test whether lag length criterion is correctly specified", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NULL,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
'"lags_criterion" has to be NaN or a character, specifying the lag length criterion.', fixed = TRUE)
})
test_that("Test whether trend is given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = NULL,
confint = 1.96,
hor = 20,
num_cores = 1),
'Please specify whether and which type of trend to include.', fixed = TRUE)
})
test_that("Test whether number of horizons is given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1,
hor = NULL,
num_cores = 1),
'Please specify the number of horizons.', fixed = TRUE)
})
test_that("Test whether lag lengt criterion is correctly given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = 'SIC',
max_lags = NaN,
trend = 0,
confint = 1,
hor = 12,
num_cores = 1),
'Possible lag length criteria are AICc, AIC or BIC. NaN if lag length is specified.', fixed = TRUE)
})
test_that("Test whether number of horizons is correctly given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1,
hor = 1.2,
num_cores = 1),
'The number of horizons has to be an integer and > 0.', fixed = TRUE)
})
test_that("Test whether trend is correctly specified", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 4,
confint = 1,
hor = 12,
num_cores = 1),
'For trend please enter 0 = no trend, 1 = trend, 2 = trend and quadratic trend.', fixed = TRUE)
})
test_that("Test whether the width of the confidence bands is correctly specified", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 1,
confint = -0.2,
hor = 12,
num_cores = 1),
'The width of the confidence bands has to be >=0.', fixed = TRUE)
})
test_that("Test whether exogenous data is a data.frame", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = as.matrix(rnorm(100)),
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 1,
confint = 1,
hor = 12,
num_cores = 1),
'Exogenous data has to be given as a data.frame.', fixed = TRUE)
})
test_that("Test whether instrument is a data.frame", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = as.matrix(rnorm(100)),
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 1,
confint = 1,
hor = 12,
num_cores = 1),
'Exogenous data has to be given as a data.frame.', fixed = TRUE)
})
test_that("Test whether lag length for exogenous data is given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = as.data.frame(rnorm(dim(endog_data)[1])),
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 1,
confint = 1,
hor = 12,
num_cores = 1),
'Please provide a lag length for the exogenous data.', fixed = TRUE)
})
test_that("Test whether lag length and lag length criterion is given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = 'AIC',
max_lags = NULL,
trend = 1,
confint = 1,
hor = 12,
num_cores = 1),
'You can not provide a lag criterion (AICc, AIC or BIC) and a fixed number of lags.
Please set lags_endog_lin to NaN if you want to use a lag length criterion.', fixed = TRUE)
})
test_that("Test that model works with automatic lag length selection", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = NaN,
exog_data = as.data.frame(rnorm(length(instrument[, 1]))),
lags_exog = 2,
contemp_data = NULL,
lags_criterion = 'AIC',
max_lags = 4,
trend = 2,
confint = 1,
hor = 12,
num_cores = 1),
NA)
})
test_that("Test that model works with automatic lag length selection AND 2SLS", {
# Set seed
set.seed(007)
# Load data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# Variable to shock with (government spending)
shock <- ag_data[sample_start:sample_end, 3]
# Generate instrument variable that is correlated with government spending
instrum <- as.data.frame(0.9*shock$Gov + rnorm(length(shock$Gov), 0, 0.02) )
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
instrum = instrum,
lags_endog_lin = NaN,
exog_data = as.data.frame(rnorm(length(instrument[, 1]))),
lags_exog = 2,
use_twosls = TRUE,
contemp_data = NULL,
lags_criterion = 'AIC',
max_lags = 4,
trend = 2,
confint = 1,
hor = 12,
num_cores = 1),
NA)
# Estimate linear model
testthat::expect_error( lp_lin_iv(endog_data,
lags_endog_lin = 4,
shock = shock,
instrum = instrum,
use_twosls = TRUE,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
nw_prewhite = T,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
NA )
})
# This example replicates results from the Supplementary Appendix
# by Ramey and Zubairy (2018), based on results from their provided
# Matlab code. The results evaluate findings from Auerbach and
# Gorodnichenko (2012) with local projections by Jordá (2005).
# The data and Matlab code is available on \url{https://www.journals.uchicago.edu/doi/10.1086/696277}{JoPE}
# Load and prepare data
# The sample length of RZ-2018 is 1948:III-2008:III
ag_data <- ag_data
sample_start <- 8
sample_end <- dim(ag_data)[1]
endog_data <- ag_data[sample_start:sample_end, 3:5]
# Choose instrument
instrument <- ag_data[sample_start:sample_end, 3]
# These results are taken from the available Matlab code by Ramey and Zubairy (2018)
# They coresspond to row 2 and 3 in table 'regg' from the Matlab output
rz_lin_results <- c(1, 1.27, 1.40, 1.46, 1.38, 1.16, 1.05, 0.90, 0.85,
0.82, 0.73, 0.53, 0.45, 0.30, 0.27, 0.28, 0.31, 0.35, 0.36, 0.27)
test_that("Compare results with RZ-2018", {
# Estimate linear model
results_lin_iv <- lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1)
lin_results <- round(results_lin_iv$irf_lin_mean[1,], 2)
testthat::expect_equal(lin_results, rz_lin_results)
})
test_that("Run use_twosls", {
# Set seed
set.seed(007)
# Load data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# Variable to shock with (government spending)
shock <- ag_data[sample_start:sample_end, 3]
# Generate instrument variable that is correlated with government spending
instrum <- as.data.frame(0.9*shock$Gov + rnorm(length(shock$Gov), 0, 0.02) )
# Estimate linear model
testthat::expect_error( lp_lin_iv(endog_data,
lags_endog_lin = 4,
shock = shock,
instrum = instrum,
use_twosls = TRUE,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
NA )
})
test_that("Test that prewhitening works", {
# Estimate linear model
testthat::expect_error(lp_lin_iv(endog_data[, 1],
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
nw_prewhite = T,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
NA)
})
test_that("Test that prewhitening and 2SLS works", {
# Set seed
set.seed(007)
# Load data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# Variable to shock with (government spending)
shock <- ag_data[sample_start:sample_end, 3]
# Generate instrument variable that is correlated with government spending
instrum <- as.data.frame(0.9*shock$Gov + rnorm(length(shock$Gov), 0, 0.02) )
# Estimate linear model
testthat::expect_error( lp_lin_iv(endog_data,
lags_endog_lin = 4,
shock = shock,
instrum = instrum,
use_twosls = TRUE,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
nw_prewhite = T,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
NA )
})
test_that("Test that 2SLS works and normal standard errors work", {
# Set seed
set.seed(007)
# Load data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# Variable to shock with (government spending)
shock <- ag_data[sample_start:sample_end, 3]
# Generate instrument variable that is correlated with government spending
instrum <- as.data.frame(0.9*shock$Gov + rnorm(length(shock$Gov), 0, 0.02) )
# Estimate linear model
testthat::expect_error( lp_lin_iv(endog_data,
lags_endog_lin = 4,
shock = shock,
instrum = instrum,
use_twosls = TRUE,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
use_nw = F,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
NA )
})
test_that("Test that cumul_multiplier work", {
# Load package data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# 'Instrument' variable
instrument <- as.data.frame(ag_data$Gov[sample_start:sample_end])
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
cumul_mult = TRUE,
lags_endog_lin = 4,
exog_data = as.data.frame(rnorm(length(instrument[, 1]))),
lags_exog = 2,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = 4,
trend = 2,
confint = 1,
hor = 12,
num_cores = 1),
NA)
})
test_that("Test cumul_mult options", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
cumul_mult = TRUE,
lags_endog_lin = NaN,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = "AIC",
max_lags = 3,
trend = 1,
confint = 1,
hor = 12,
num_cores = 1),
'The option cumul_mult = TRUE only works for a fixed number of lags.', fixed = TRUE)
})
test_that("Test whether 2SLS with cumul_mult works", {
shock <- ag_data[sample_start:sample_end, 3]
exog_tbl <- tibble(x1 = rnorm(dim(shock)[1]), x2 = rnorm(dim(shock)[1]))
testthat::expect_error(lp_lin_iv(
endog_data = endog_data,
shock = shock,
cumul_mult = T,
instr = NULL,
instrum = instrument,
use_twosls = TRUE,
lags_endog_lin = 4,
exog_data = exog_tbl,
lags_exog = 4,
lags_criterion = NaN,
trend = 0,
confint = 1.96,
hor = 12,
NA)
)})
AdaemmerP/lpirfs documentation built on July 27, 2023, 6:15 p.m.
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
# Load package data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# 'Instrument' variable
instrument <- as.data.frame(ag_data$Gov[sample_start:sample_end])
test_that("Test whether data is a data.frame", {
endog_data <- as.matrix(endog_data)
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
'The data has to be a data.frame().', fixed = TRUE)
})
test_that("Test whether instrument is given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = NULL,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
'You have to provide an instrument to shock with.', fixed = TRUE)
})
test_that("Test whether lags_endog_lin is falsely NaN", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = NaN,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
'"lags_endog_lin" can only be NaN if a lag length criterion is given.', fixed = TRUE)
})
test_that("Test whether lag length criterion is correctly specified", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NULL,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
'"lags_criterion" has to be NaN or a character, specifying the lag length criterion.', fixed = TRUE)
})
test_that("Test whether trend is given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = NULL,
confint = 1.96,
hor = 20,
num_cores = 1),
'Please specify whether and which type of trend to include.', fixed = TRUE)
})
test_that("Test whether number of horizons is given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1,
hor = NULL,
num_cores = 1),
'Please specify the number of horizons.', fixed = TRUE)
})
test_that("Test whether lag lengt criterion is correctly given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = 'SIC',
max_lags = NaN,
trend = 0,
confint = 1,
hor = 12,
num_cores = 1),
'Possible lag length criteria are AICc, AIC or BIC. NaN if lag length is specified.', fixed = TRUE)
})
test_that("Test whether number of horizons is correctly given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1,
hor = 1.2,
num_cores = 1),
'The number of horizons has to be an integer and > 0.', fixed = TRUE)
})
test_that("Test whether trend is correctly specified", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 4,
confint = 1,
hor = 12,
num_cores = 1),
'For trend please enter 0 = no trend, 1 = trend, 2 = trend and quadratic trend.', fixed = TRUE)
})
test_that("Test whether the width of the confidence bands is correctly specified", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 1,
confint = -0.2,
hor = 12,
num_cores = 1),
'The width of the confidence bands has to be >=0.', fixed = TRUE)
})
test_that("Test whether exogenous data is a data.frame", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = as.matrix(rnorm(100)),
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 1,
confint = 1,
hor = 12,
num_cores = 1),
'Exogenous data has to be given as a data.frame.', fixed = TRUE)
})
test_that("Test whether instrument is a data.frame", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = as.matrix(rnorm(100)),
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 1,
confint = 1,
hor = 12,
num_cores = 1),
'Exogenous data has to be given as a data.frame.', fixed = TRUE)
})
test_that("Test whether lag length for exogenous data is given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = as.data.frame(rnorm(dim(endog_data)[1])),
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 1,
confint = 1,
hor = 12,
num_cores = 1),
'Please provide a lag length for the exogenous data.', fixed = TRUE)
})
test_that("Test whether lag length and lag length criterion is given", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = 'AIC',
max_lags = NULL,
trend = 1,
confint = 1,
hor = 12,
num_cores = 1),
'You can not provide a lag criterion (AICc, AIC or BIC) and a fixed number of lags.
Please set lags_endog_lin to NaN if you want to use a lag length criterion.', fixed = TRUE)
})
test_that("Test that model works with automatic lag length selection", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = NaN,
exog_data = as.data.frame(rnorm(length(instrument[, 1]))),
lags_exog = 2,
contemp_data = NULL,
lags_criterion = 'AIC',
max_lags = 4,
trend = 2,
confint = 1,
hor = 12,
num_cores = 1),
NA)
})
test_that("Test that model works with automatic lag length selection AND 2SLS", {
# Set seed
set.seed(007)
# Load data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# Variable to shock with (government spending)
shock <- ag_data[sample_start:sample_end, 3]
# Generate instrument variable that is correlated with government spending
instrum <- as.data.frame(0.9*shock$Gov + rnorm(length(shock$Gov), 0, 0.02) )
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
instrum = instrum,
lags_endog_lin = NaN,
exog_data = as.data.frame(rnorm(length(instrument[, 1]))),
lags_exog = 2,
use_twosls = TRUE,
contemp_data = NULL,
lags_criterion = 'AIC',
max_lags = 4,
trend = 2,
confint = 1,
hor = 12,
num_cores = 1),
NA)
# Estimate linear model
testthat::expect_error( lp_lin_iv(endog_data,
lags_endog_lin = 4,
shock = shock,
instrum = instrum,
use_twosls = TRUE,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
nw_prewhite = T,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
NA )
})
# This example replicates results from the Supplementary Appendix
# by Ramey and Zubairy (2018), based on results from their provided
# Matlab code. The results evaluate findings from Auerbach and
# Gorodnichenko (2012) with local projections by Jordá (2005).
# The data and Matlab code is available on \url{https://www.journals.uchicago.edu/doi/10.1086/696277}{JoPE}
# Load and prepare data
# The sample length of RZ-2018 is 1948:III-2008:III
ag_data <- ag_data
sample_start <- 8
sample_end <- dim(ag_data)[1]
endog_data <- ag_data[sample_start:sample_end, 3:5]
# Choose instrument
instrument <- ag_data[sample_start:sample_end, 3]
# These results are taken from the available Matlab code by Ramey and Zubairy (2018)
# They coresspond to row 2 and 3 in table 'regg' from the Matlab output
rz_lin_results <- c(1, 1.27, 1.40, 1.46, 1.38, 1.16, 1.05, 0.90, 0.85,
0.82, 0.73, 0.53, 0.45, 0.30, 0.27, 0.28, 0.31, 0.35, 0.36, 0.27)
test_that("Compare results with RZ-2018", {
# Estimate linear model
results_lin_iv <- lp_lin_iv(endog_data,
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1)
lin_results <- round(results_lin_iv$irf_lin_mean[1,], 2)
testthat::expect_equal(lin_results, rz_lin_results)
})
test_that("Run use_twosls", {
# Set seed
set.seed(007)
# Load data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# Variable to shock with (government spending)
shock <- ag_data[sample_start:sample_end, 3]
# Generate instrument variable that is correlated with government spending
instrum <- as.data.frame(0.9*shock$Gov + rnorm(length(shock$Gov), 0, 0.02) )
# Estimate linear model
testthat::expect_error( lp_lin_iv(endog_data,
lags_endog_lin = 4,
shock = shock,
instrum = instrum,
use_twosls = TRUE,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
NA )
})
test_that("Test that prewhitening works", {
# Estimate linear model
testthat::expect_error(lp_lin_iv(endog_data[, 1],
shock = instrument,
lags_endog_lin = 4,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
nw_prewhite = T,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
NA)
})
test_that("Test that prewhitening and 2SLS works", {
# Set seed
set.seed(007)
# Load data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# Variable to shock with (government spending)
shock <- ag_data[sample_start:sample_end, 3]
# Generate instrument variable that is correlated with government spending
instrum <- as.data.frame(0.9*shock$Gov + rnorm(length(shock$Gov), 0, 0.02) )
# Estimate linear model
testthat::expect_error( lp_lin_iv(endog_data,
lags_endog_lin = 4,
shock = shock,
instrum = instrum,
use_twosls = TRUE,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
nw_prewhite = T,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
NA )
})
test_that("Test that 2SLS works and normal standard errors work", {
# Set seed
set.seed(007)
# Load data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# Variable to shock with (government spending)
shock <- ag_data[sample_start:sample_end, 3]
# Generate instrument variable that is correlated with government spending
instrum <- as.data.frame(0.9*shock$Gov + rnorm(length(shock$Gov), 0, 0.02) )
# Estimate linear model
testthat::expect_error( lp_lin_iv(endog_data,
lags_endog_lin = 4,
shock = shock,
instrum = instrum,
use_twosls = TRUE,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = NaN,
use_nw = F,
trend = 0,
confint = 1.96,
hor = 20,
num_cores = 1),
NA )
})
test_that("Test that cumul_multiplier work", {
# Load package data
ag_data <- ag_data
sample_start <- 7
sample_end <- dim(ag_data)[1]
# Endogenous data
endog_data <- ag_data[sample_start:sample_end,3:5]
# 'Instrument' variable
instrument <- as.data.frame(ag_data$Gov[sample_start:sample_end])
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
cumul_mult = TRUE,
lags_endog_lin = 4,
exog_data = as.data.frame(rnorm(length(instrument[, 1]))),
lags_exog = 2,
contemp_data = NULL,
lags_criterion = NaN,
max_lags = 4,
trend = 2,
confint = 1,
hor = 12,
num_cores = 1),
NA)
})
test_that("Test cumul_mult options", {
testthat::expect_error(lp_lin_iv(endog_data,
shock = instrument,
cumul_mult = TRUE,
lags_endog_lin = NaN,
exog_data = NULL,
lags_exog = NULL,
contemp_data = NULL,
lags_criterion = "AIC",
max_lags = 3,
trend = 1,
confint = 1,
hor = 12,
num_cores = 1),
'The option cumul_mult = TRUE only works for a fixed number of lags.', fixed = TRUE)
})
test_that("Test whether 2SLS with cumul_mult works", {
shock <- ag_data[sample_start:sample_end, 3]
exog_tbl <- tibble(x1 = rnorm(dim(shock)[1]), x2 = rnorm(dim(shock)[1]))
testthat::expect_error(lp_lin_iv(
endog_data = endog_data,
shock = shock,
cumul_mult = T,
instr = NULL,
instrum = instrument,
use_twosls = TRUE,
lags_endog_lin = 4,
exog_data = exog_tbl,
lags_exog = 4,
lags_criterion = NaN,
trend = 0,
confint = 1.96,
hor = 12,
NA)
)})
R Package Documentation
Browse R Packages
We want your feedback!
Note that we can't provide technical support on individual packages. You should contact the package authors for that.
Embedding an R snippet on your website
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.