tests/testthat/test-lp_lin_panel.R
In AdaemmerP/lpirfs: Local Projections Impulse Response Functions
library(dplyr)
set.seed(123)
# Simulate panel data
N <- 10
TS <- 30
cross_section <- sort(LETTERS[rep(seq(1,N, 1), TS)])
time_section <- rep(seq(1,TS, 1), N)
data_set <- tibble(cross_section, time_section) %>%
group_by(cross_section) %>%
mutate(x_2 = rnorm(TS)) %>%
mutate(x_1 = rnorm(TS)) %>%
mutate(x_3 = rnorm(TS)) %>%
mutate(x_4 = rnorm(TS)) %>%
mutate(y = 0.3*x_1 + 0.4*x_2 + 0.5*x_3 + 0.6*x_4 + rnorm(TS)) %>%
ungroup()
test_that("Test that column names are correct", {
names(data_set)[3:4] <- c("cross_id", "date_id")
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You cannot use the column names 'cross_id' or 'date_id' besides the first two columns of your data.frame.
Please rename them.", fixed = TRUE)
})
test_that("Test that column names are correct", {
names(data_set)[3] <- "Test_lag_name"
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
'Please do not use column names that include the string "lag_" in the name.
This cause later naming problems', fixed = TRUE)
})
test_that("Test that data frame is given", {
testthat::expect_error(lp_lin_panel(data_set = NULL,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You have to provide the panel data set.", fixed = TRUE)
})
test_that("Test that name of endogenous variable is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = NULL,
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You have to provide the name of the endogenous variable.", fixed = TRUE)
})
test_that("Test that name of shock variable is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = NULL,
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You have to provide the name of the variable to shock with.", fixed = TRUE)
})
test_that("Test whether name of instrument variable is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = TRUE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You have to provide the name of the instrument.", fixed = TRUE)
})
test_that("Test whether panel model type is correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "soso",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"The type of the panel model has to be 'within', 'random', 'ht', 'between', 'pooling' or 'fd'. See
the vignette of the plm package for details.", fixed = TRUE)
})
test_that("Test whether panel effect is correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "soso",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"The effect introduced in the model has to be 'individual', 'time', 'twoways' or 'nested'.
See the vignette of the plm package for details.", fixed = TRUE)
})
test_that("Test whether lag lengths for the exogenous data is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = NULL,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You have to provide the lag lengths for the exogenous data with lagged impact.", fixed = TRUE)
})
test_that("Test whether lag lengths for the exogenous data (first differences) is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = NULL,
confint = 1.67,
hor = 10),
"You have to provide the lag lengths for the exogenous data with lagged impact of first differences.", fixed = TRUE)
})
test_that("Test that width of confidence bands is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = NULL,
hor = 10),
'Please specify a value for the width of the confidence bands.', fixed = TRUE)
})
test_that("Test that width of confidence bands is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = -1,
hor = 10),
'The width of the confidence bands has to be >=0.', fixed = TRUE)
})
test_that("Test that robust covariance matrix is correctly given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = "soso",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = 10),
"The choices for robust covariance estimation are 'vcovBK', 'vcovDC', 'vcovHC', 'vcovNW', 'vcovSCC' and 'Vcxt'.
For details, see the vignette of the plm package and Miller (2017).", fixed = TRUE)
})
test_that("Test that horizon integer is correctly specified", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = -1.1),
'The number of horizons has to be an integer and > 0.', fixed = TRUE)
})
test_that("Test that gmm options are correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
use_gmm = TRUE,
gmm_effect = "soso",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = 5),
'The effect for gmm has to be "twoways" (default) or "individual".', fixed = TRUE)
})
test_that("Test that gmm options are correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
use_gmm = TRUE,
gmm_transformation = "soso",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = 5),
'The transformation to apply to the model has to either be "d" (default)
for the "difference GMM" model or "ld" for the "system GMM".', fixed = TRUE)
})
test_that("Test that gmm options are correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = "vcovSCC",
use_gmm = TRUE,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = 5),
'If you want to estimate a gmm model, set "robust_cov = NULL".', fixed = TRUE)
})
test_that("Test that gmm options are correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
use_gmm = TRUE,
gmm_effect = "soso",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = 5),
'The effect for gmm has to be "twoways" (default) or "individual".', fixed = TRUE)
})
# Estimate panel model
results_panel <- suppressWarnings(
lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10))
lp_lin_panel_results <- results_panel$reg_summaries[[1]]
yx_data <- results_panel$xy_data_sets[[1]]
x_names <- colnames(yx_data)[!colnames(yx_data) %in% c("cross_id", "date_id", "y")]
panel_formula <- as.formula(paste("y~", paste(x_names, collapse = "+")))
panel_results <- plm::plm(formula = panel_formula,
data = yx_data,
index = c("cross_id", "date_id"),
model = "within",
effect = "individual")
test_that("Test whether output coefficients from 'lp_lin_panel' coincide
with coefficients estimated by using the matrix prepared in 'lp-panel_lin'", {
testthat::expect_equal(panel_results$coefficients, results_panel$reg_summaries[[1]]$coefficients[,1])
})
test_that("Test that model throws no error when estimating robust
covariance matrix", {
# All robust cov matrices except of 'vcovDC'
cov_mat <- c('vcovBK', 'vcovHC', 'vcovNW', 'vcovSCC')
for (ii in seq_along(cov_mat)){
# Estimate panel model
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = cov_mat[ii],
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 2),
NA)
}
# Estimate panel model with 'vcovDC'
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "twoways",
robust_cov = 'vcovDC',
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 2),
NA)
# Test 'vcovSCC' with different type
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = 'vcovSCC',
robust_type = "HC1",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 2),
NA)
})
test_that("Test that data is correctly computed.", {
# Estimate panel model
results_lin_panel <- lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = FALSE,
shock = "x_1",
diff_shock = TRUE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = "vcovSCC",
c_exog_data = c("x_2", "x_3", "x_4"), #colnames(data_set)[4:6],
l_exog_data = c("x_2", "x_3", "x_4"),
lags_exog_data = 2,
c_fd_exog_data = c("x_2", "x_3", "x_4"),
l_fd_exog_data = c("x_2", "x_3", "x_4"),
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10)
# Save data set for h = 0
data_output <- results_lin_panel$xy_data_sets[[1]]
# Save contemporaneous output data, h = 0
c_exog_data_output <- results_lin_panel$xy_data_sets[[1]][, c("x_2", "x_3", "x_4")]
# Save lagged exogenous data, h = 0
lag_names_1 <- paste(c("x_2", "x_3", "x_4"), c("lag_1"), sep = "_")
lag_names_2 <- paste(c("x_2", "x_3", "x_4"), c("lag_2"), sep = "_")
lag_names <- c(lag_names_1, lag_names_2)
l_exog_data_output <- data_output[, which(names(data_output) %in% lag_names)]
# Saved lagged data of first differences, h = 0
dlag_names_1 <- paste(c("dx_2", "dx_3", "dx_4"), c("lag_1"), sep = "_")
dlag_names_2 <- paste(c("dx_2", "dx_3", "dx_4"), c("lag_2"), sep = "_")
dlag_names <- c(dlag_names_1, dlag_names_2)
dl_exog_data_output <- data_output[, which(names(data_output) %in% dlag_names)]
# Compute manually contemporaneous data
# Use lead = 3, because lag length = 2 AND lags of first differences = 2.
c_exog_data_manual <- data_set %>%
dplyr::group_by(cross_section) %>%
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~dplyr::lead(., 3))) %>%
dplyr::ungroup() %>%
na.omit() %>%
dplyr::select(-x_1, -y, -cross_section, -time_section)
dim_c_exog_data_manual <- paste(dim(c_exog_data_manual), collapse = "")
dim_c_exog_data_output <- paste(dim(c_exog_data_output), collapse = "")
# Compare contemporaneous data
# testthat::expect_equal(c_exog_data_manual, c_exog_data_output) Deprecated as of dplyr 1.0.0
testthat::expect_equal(dim_c_exog_data_manual, dim_c_exog_data_output)
# Function which takes first differences and sets the first value to NA
# to be consistent with dplyr
diff_function <- function(data){
return(c(NA, diff(data)))
}
# Compute manually lagged exogenous data
# First lag
l_1_exog_data_manual <- data_set %>%
dplyr::group_by(cross_section) %>%
dplyr::mutate_at(vars(y), list(~diff_function(.))) %>% # This constructs first differences
dplyr::mutate_at(vars(y), list(~dplyr::lag(., 2))) %>% # This accounts for the lag = 2 of first differences
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~dplyr::lag(., 1))) %>% # This is the first lag of exogenous data
dplyr::ungroup() %>%
na.omit() %>%
dplyr::select(-x_1, -y, -cross_section, -time_section) %>%
dplyr::rename(x_2_lag_1 = x_2,
x_3_lag_1 = x_3,
x_4_lag_1 = x_4)
# Second lag
l_2_exog_data_manual <- data_set %>%
dplyr::group_by(cross_section) %>%
dplyr::mutate_at(vars(y), list(~diff_function(.))) %>% # This constructs first differences
dplyr::mutate_at(vars(y), list(~dplyr::lag(., 2))) %>% # This accounts for the lag = 2 of first differences
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~dplyr::lag(., 2))) %>% # This is the second lag of exogenous data
dplyr::ungroup() %>%
na.omit() %>%
dplyr::select(-x_1, -y, -cross_section, -time_section) %>%
dplyr::rename(x_2_lag_2 = x_2,
x_3_lag_2 = x_3,
x_4_lag_2 = x_4)
l_exog_data_manual <- cbind(l_1_exog_data_manual, l_2_exog_data_manual)
dim_l_exog_data_output <- paste(dim(l_exog_data_output), collapse = "")
dim_l_exog_data_manual <- paste(dim(l_exog_data_manual), collapse = "")
# Compare contemporaneous data
# testthat::expect_equal(l_exog_data_output, l_exog_data_manual) Deprecated as of dplyr 1.0.0
testthat::expect_equal(dim_l_exog_data_output, dim_l_exog_data_manual)
# Compute lagged exogenous data of first differences
# First lag
dl_1_exog_data_manual <- data_set %>%
dplyr::group_by(cross_section) %>%
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~diff_function(.))) %>% # This constructs first differences
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~dplyr::lag(., 1))) %>% # Take first lag of first differences
dplyr::ungroup() %>%
dplyr::select(-x_1, -y, -cross_section, -time_section) %>%
dplyr::rename(dx_2_lag_1 = x_2,
dx_3_lag_1 = x_3,
dx_4_lag_1 = x_4)
# Second lag
dl_2_exog_data_manual <- data_set %>%
dplyr::group_by(cross_section) %>%
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~diff_function(.))) %>% # This constructs first differences
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~dplyr::lag(., 2))) %>% # Take second lag of first differences
dplyr::ungroup() %>%
dplyr::select(-x_1, -y, -cross_section, -time_section) %>%
dplyr::rename(dx_2_lag_2 = x_2,
dx_3_lag_2 = x_3,
dx_4_lag_2 = x_4)
dl_exog_data_manual <- na.omit(cbind(dl_1_exog_data_manual, dl_2_exog_data_manual)) %>%
as_tibble()
# Compare lagged exogenous data of first differences
# Use rowSums due to different order of manual and output data
testthat::expect_equal(rowSums(dl_exog_data_manual), rowSums(dl_exog_data_output))
})
test_that("Test model when estimating normal multipliers", {
# Estimate panel model
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = FALSE,
shock = "x_1",
diff_shock = TRUE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = "vcovSCC",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
NA)
})
test_that("Test gmm approach", {
set.seed(123)
# Simulate panel data with large cross dimension
N <- 140
TS <- 8
cross_section <- sort(rep(seq(1,N, 1), TS))
time_section <- rep(seq(1,TS, 1), N)
data_set <- tibble(cross_section, time_section) %>%
group_by(cross_section) %>%
mutate(x_1 = rnorm(TS)) %>%
mutate(x_2 = rnorm(TS)) %>%
mutate(x_3 = rnorm(TS)) %>%
mutate(x_4 = rnorm(TS)) %>%
mutate(y = 0.3*x_1 + 0.4*x_2 + 0.5*x_3 + 0.6*x_4 + rnorm(TS)) %>%
ungroup()
# Estimate panel model
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
use_gmm = TRUE,
c_exog_data = NULL,
l_exog_data = "y",
lags_exog_data = 1,
c_fd_exog_data = NULL,
l_fd_exog_data = NULL,
lags_fd_exog_data = NULL,
confint = 1.67,
hor = 3),
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.
library(dplyr)
set.seed(123)
# Simulate panel data
N <- 10
TS <- 30
cross_section <- sort(LETTERS[rep(seq(1,N, 1), TS)])
time_section <- rep(seq(1,TS, 1), N)
data_set <- tibble(cross_section, time_section) %>%
group_by(cross_section) %>%
mutate(x_2 = rnorm(TS)) %>%
mutate(x_1 = rnorm(TS)) %>%
mutate(x_3 = rnorm(TS)) %>%
mutate(x_4 = rnorm(TS)) %>%
mutate(y = 0.3*x_1 + 0.4*x_2 + 0.5*x_3 + 0.6*x_4 + rnorm(TS)) %>%
ungroup()
test_that("Test that column names are correct", {
names(data_set)[3:4] <- c("cross_id", "date_id")
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You cannot use the column names 'cross_id' or 'date_id' besides the first two columns of your data.frame.
Please rename them.", fixed = TRUE)
})
test_that("Test that column names are correct", {
names(data_set)[3] <- "Test_lag_name"
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
'Please do not use column names that include the string "lag_" in the name.
This cause later naming problems', fixed = TRUE)
})
test_that("Test that data frame is given", {
testthat::expect_error(lp_lin_panel(data_set = NULL,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You have to provide the panel data set.", fixed = TRUE)
})
test_that("Test that name of endogenous variable is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = NULL,
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You have to provide the name of the endogenous variable.", fixed = TRUE)
})
test_that("Test that name of shock variable is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = NULL,
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You have to provide the name of the variable to shock with.", fixed = TRUE)
})
test_that("Test whether name of instrument variable is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = TRUE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You have to provide the name of the instrument.", fixed = TRUE)
})
test_that("Test whether panel model type is correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "soso",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"The type of the panel model has to be 'within', 'random', 'ht', 'between', 'pooling' or 'fd'. See
the vignette of the plm package for details.", fixed = TRUE)
})
test_that("Test whether panel effect is correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "soso",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"The effect introduced in the model has to be 'individual', 'time', 'twoways' or 'nested'.
See the vignette of the plm package for details.", fixed = TRUE)
})
test_that("Test whether lag lengths for the exogenous data is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = NULL,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
"You have to provide the lag lengths for the exogenous data with lagged impact.", fixed = TRUE)
})
test_that("Test whether lag lengths for the exogenous data (first differences) is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = NULL,
confint = 1.67,
hor = 10),
"You have to provide the lag lengths for the exogenous data with lagged impact of first differences.", fixed = TRUE)
})
test_that("Test that width of confidence bands is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = NULL,
hor = 10),
'Please specify a value for the width of the confidence bands.', fixed = TRUE)
})
test_that("Test that width of confidence bands is given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = -1,
hor = 10),
'The width of the confidence bands has to be >=0.', fixed = TRUE)
})
test_that("Test that robust covariance matrix is correctly given", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = "soso",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = 10),
"The choices for robust covariance estimation are 'vcovBK', 'vcovDC', 'vcovHC', 'vcovNW', 'vcovSCC' and 'Vcxt'.
For details, see the vignette of the plm package and Miller (2017).", fixed = TRUE)
})
test_that("Test that horizon integer is correctly specified", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = -1.1),
'The number of horizons has to be an integer and > 0.', fixed = TRUE)
})
test_that("Test that gmm options are correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
use_gmm = TRUE,
gmm_effect = "soso",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = 5),
'The effect for gmm has to be "twoways" (default) or "individual".', fixed = TRUE)
})
test_that("Test that gmm options are correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
use_gmm = TRUE,
gmm_transformation = "soso",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = 5),
'The transformation to apply to the model has to either be "d" (default)
for the "difference GMM" model or "ld" for the "system GMM".', fixed = TRUE)
})
test_that("Test that gmm options are correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = "vcovSCC",
use_gmm = TRUE,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = 5),
'If you want to estimate a gmm model, set "robust_cov = NULL".', fixed = TRUE)
})
test_that("Test that gmm options are correct", {
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
use_gmm = TRUE,
gmm_effect = "soso",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 1,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1,
hor = 5),
'The effect for gmm has to be "twoways" (default) or "individual".', fixed = TRUE)
})
# Estimate panel model
results_panel <- suppressWarnings(
lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10))
lp_lin_panel_results <- results_panel$reg_summaries[[1]]
yx_data <- results_panel$xy_data_sets[[1]]
x_names <- colnames(yx_data)[!colnames(yx_data) %in% c("cross_id", "date_id", "y")]
panel_formula <- as.formula(paste("y~", paste(x_names, collapse = "+")))
panel_results <- plm::plm(formula = panel_formula,
data = yx_data,
index = c("cross_id", "date_id"),
model = "within",
effect = "individual")
test_that("Test whether output coefficients from 'lp_lin_panel' coincide
with coefficients estimated by using the matrix prepared in 'lp-panel_lin'", {
testthat::expect_equal(panel_results$coefficients, results_panel$reg_summaries[[1]]$coefficients[,1])
})
test_that("Test that model throws no error when estimating robust
covariance matrix", {
# All robust cov matrices except of 'vcovDC'
cov_mat <- c('vcovBK', 'vcovHC', 'vcovNW', 'vcovSCC')
for (ii in seq_along(cov_mat)){
# Estimate panel model
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = cov_mat[ii],
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 2),
NA)
}
# Estimate panel model with 'vcovDC'
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "twoways",
robust_cov = 'vcovDC',
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 2),
NA)
# Test 'vcovSCC' with different type
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = 'vcovSCC',
robust_type = "HC1",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 2),
NA)
})
test_that("Test that data is correctly computed.", {
# Estimate panel model
results_lin_panel <- lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = FALSE,
shock = "x_1",
diff_shock = TRUE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = "vcovSCC",
c_exog_data = c("x_2", "x_3", "x_4"), #colnames(data_set)[4:6],
l_exog_data = c("x_2", "x_3", "x_4"),
lags_exog_data = 2,
c_fd_exog_data = c("x_2", "x_3", "x_4"),
l_fd_exog_data = c("x_2", "x_3", "x_4"),
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10)
# Save data set for h = 0
data_output <- results_lin_panel$xy_data_sets[[1]]
# Save contemporaneous output data, h = 0
c_exog_data_output <- results_lin_panel$xy_data_sets[[1]][, c("x_2", "x_3", "x_4")]
# Save lagged exogenous data, h = 0
lag_names_1 <- paste(c("x_2", "x_3", "x_4"), c("lag_1"), sep = "_")
lag_names_2 <- paste(c("x_2", "x_3", "x_4"), c("lag_2"), sep = "_")
lag_names <- c(lag_names_1, lag_names_2)
l_exog_data_output <- data_output[, which(names(data_output) %in% lag_names)]
# Saved lagged data of first differences, h = 0
dlag_names_1 <- paste(c("dx_2", "dx_3", "dx_4"), c("lag_1"), sep = "_")
dlag_names_2 <- paste(c("dx_2", "dx_3", "dx_4"), c("lag_2"), sep = "_")
dlag_names <- c(dlag_names_1, dlag_names_2)
dl_exog_data_output <- data_output[, which(names(data_output) %in% dlag_names)]
# Compute manually contemporaneous data
# Use lead = 3, because lag length = 2 AND lags of first differences = 2.
c_exog_data_manual <- data_set %>%
dplyr::group_by(cross_section) %>%
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~dplyr::lead(., 3))) %>%
dplyr::ungroup() %>%
na.omit() %>%
dplyr::select(-x_1, -y, -cross_section, -time_section)
dim_c_exog_data_manual <- paste(dim(c_exog_data_manual), collapse = "")
dim_c_exog_data_output <- paste(dim(c_exog_data_output), collapse = "")
# Compare contemporaneous data
# testthat::expect_equal(c_exog_data_manual, c_exog_data_output) Deprecated as of dplyr 1.0.0
testthat::expect_equal(dim_c_exog_data_manual, dim_c_exog_data_output)
# Function which takes first differences and sets the first value to NA
# to be consistent with dplyr
diff_function <- function(data){
return(c(NA, diff(data)))
}
# Compute manually lagged exogenous data
# First lag
l_1_exog_data_manual <- data_set %>%
dplyr::group_by(cross_section) %>%
dplyr::mutate_at(vars(y), list(~diff_function(.))) %>% # This constructs first differences
dplyr::mutate_at(vars(y), list(~dplyr::lag(., 2))) %>% # This accounts for the lag = 2 of first differences
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~dplyr::lag(., 1))) %>% # This is the first lag of exogenous data
dplyr::ungroup() %>%
na.omit() %>%
dplyr::select(-x_1, -y, -cross_section, -time_section) %>%
dplyr::rename(x_2_lag_1 = x_2,
x_3_lag_1 = x_3,
x_4_lag_1 = x_4)
# Second lag
l_2_exog_data_manual <- data_set %>%
dplyr::group_by(cross_section) %>%
dplyr::mutate_at(vars(y), list(~diff_function(.))) %>% # This constructs first differences
dplyr::mutate_at(vars(y), list(~dplyr::lag(., 2))) %>% # This accounts for the lag = 2 of first differences
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~dplyr::lag(., 2))) %>% # This is the second lag of exogenous data
dplyr::ungroup() %>%
na.omit() %>%
dplyr::select(-x_1, -y, -cross_section, -time_section) %>%
dplyr::rename(x_2_lag_2 = x_2,
x_3_lag_2 = x_3,
x_4_lag_2 = x_4)
l_exog_data_manual <- cbind(l_1_exog_data_manual, l_2_exog_data_manual)
dim_l_exog_data_output <- paste(dim(l_exog_data_output), collapse = "")
dim_l_exog_data_manual <- paste(dim(l_exog_data_manual), collapse = "")
# Compare contemporaneous data
# testthat::expect_equal(l_exog_data_output, l_exog_data_manual) Deprecated as of dplyr 1.0.0
testthat::expect_equal(dim_l_exog_data_output, dim_l_exog_data_manual)
# Compute lagged exogenous data of first differences
# First lag
dl_1_exog_data_manual <- data_set %>%
dplyr::group_by(cross_section) %>%
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~diff_function(.))) %>% # This constructs first differences
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~dplyr::lag(., 1))) %>% # Take first lag of first differences
dplyr::ungroup() %>%
dplyr::select(-x_1, -y, -cross_section, -time_section) %>%
dplyr::rename(dx_2_lag_1 = x_2,
dx_3_lag_1 = x_3,
dx_4_lag_1 = x_4)
# Second lag
dl_2_exog_data_manual <- data_set %>%
dplyr::group_by(cross_section) %>%
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~diff_function(.))) %>% # This constructs first differences
dplyr::mutate_at(vars(x_2, x_3, x_4), list(~dplyr::lag(., 2))) %>% # Take second lag of first differences
dplyr::ungroup() %>%
dplyr::select(-x_1, -y, -cross_section, -time_section) %>%
dplyr::rename(dx_2_lag_2 = x_2,
dx_3_lag_2 = x_3,
dx_4_lag_2 = x_4)
dl_exog_data_manual <- na.omit(cbind(dl_1_exog_data_manual, dl_2_exog_data_manual)) %>%
as_tibble()
# Compare lagged exogenous data of first differences
# Use rowSums due to different order of manual and output data
testthat::expect_equal(rowSums(dl_exog_data_manual), rowSums(dl_exog_data_output))
})
test_that("Test model when estimating normal multipliers", {
# Estimate panel model
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = FALSE,
shock = "x_1",
diff_shock = TRUE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = "vcovSCC",
c_exog_data = colnames(data_set)[4:6],
l_exog_data = colnames(data_set)[4:6],
lags_exog_data = 2,
c_fd_exog_data = colnames(data_set)[4:6],
l_fd_exog_data = colnames(data_set)[4:6],
lags_fd_exog_data = 2,
confint = 1.67,
hor = 10),
NA)
})
test_that("Test gmm approach", {
set.seed(123)
# Simulate panel data with large cross dimension
N <- 140
TS <- 8
cross_section <- sort(rep(seq(1,N, 1), TS))
time_section <- rep(seq(1,TS, 1), N)
data_set <- tibble(cross_section, time_section) %>%
group_by(cross_section) %>%
mutate(x_1 = rnorm(TS)) %>%
mutate(x_2 = rnorm(TS)) %>%
mutate(x_3 = rnorm(TS)) %>%
mutate(x_4 = rnorm(TS)) %>%
mutate(y = 0.3*x_1 + 0.4*x_2 + 0.5*x_3 + 0.6*x_4 + rnorm(TS)) %>%
ungroup()
# Estimate panel model
testthat::expect_error(lp_lin_panel(data_set = data_set,
data_sample = 'Full',
endog_data = "y",
cumul_mult = TRUE,
shock = "x_1",
diff_shock = FALSE,
iv_reg = FALSE,
instrum = NULL,
panel_model = "within",
panel_effect = "individual",
robust_cov = NULL,
use_gmm = TRUE,
c_exog_data = NULL,
l_exog_data = "y",
lags_exog_data = 1,
c_fd_exog_data = NULL,
l_fd_exog_data = NULL,
lags_fd_exog_data = NULL,
confint = 1.67,
hor = 3),
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.