inst/scripts/ff92_cross_sections.R
In GregoryBrownson/EAPR: Empirical Asset Pricing Research Platform
#rm(list = ls())
#library(data.table)
library(devtools)
library(ggplot2)
library(gridExtra)
library(lmtest)
library(purrr)
library(RobStatTM)
library(sandwich)
library(tables)
library(xts)
library(EAPR)
devtools::load_all(".")
eapr <- EAPR::extract("gsb25")
dat.filtered <- EAPR::filter.ff92(eapr)
size.cs <- EAPR::fama_macbeth(adj_ret * 100 ~ log_size, x = dat.filtered)
## SIZE
# LS and Robust regressions
fit.classic.size <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "log_size"))
coef(lm(adj_ret * 100 ~ log_size, data = x))
})
fit.robust.size <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "log_size"))
control <- lmrobdet.control(efficiency = 0.99)
unname(coef(lmrobdetMM(adj_ret * 100 ~ log_size, data = x, control = control)))
})
dates <- as.Date(names(fit.classic.size))
fit.classic.size.dt <- cbind(data.table(date = dates), data.table(Reduce(rbind, fit.classic.size)))
colnames(fit.classic.size.dt) <- c("beta_0", "beta_1")
fit.classic.size.dt <- cbind(data.table(date = dates), fit.classic.size.dt)
fit.robust.size.dt <- data.table(matrix(unlist(fit.robust.size), ncol = 2, byrow = TRUE))
colnames(fit.robust.size.dt) <- c("beta_0", "beta_1")
fit.robust.size.dt <- cbind(data.table(date = dates), fit.robust.size.dt)
fit.size.dt <- data.table(date = dates, LS = fit.classic.size.dt$beta_1, ROBUST = fit.robust.size.dt$beta_1)
fit.size.dt.melted <- melt(fit.size.dt, id.vars = "date")
fit.size.jan.dt.melted <- fit.size.dt.melted[month(date) == 1]
y.lim <- ceiling(max(abs(fit.classic.size.dt$beta_1)))
jpeg("./inst/etc/img/LS_size_slopes_1963_2018.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.classic.size.dt, aes(x = date, y = beta_1), col = "dodgerblue2") +
geom_point(data = fit.classic.size.dt[month(date) == 1], aes(x = date, y = beta_1), size = 1) +
geom_hline(yintercept = 0) +
ggtitle("Time Series of LS Size Slopes 1963-2018") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank())
dev.off()
y.lim <- ceiling(max(abs(fit.size.dt.melted[year(date) <= 1990]$value)))
jpeg("./inst/etc/img/size_slopes_1963_1990.jpeg", width = 700, height = 500)
ggplot() +
geom_hline(yintercept = 0) +
geom_line(data = fit.size.dt.melted[year(date) <= 1990], aes(x = date, y = value), col = "dodgerblue2") +
geom_point(data = fit.size.jan.dt.melted[year(date) <= 1990], aes(x = date, y = value), size = 1) +
facet_grid(rows = vars(variable), switch = "both") +
ggtitle("Time Series of Size Slopes 1963-1990") +
xlab("Date") +
scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "5 years", date_minor_breaks = "1 year", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank()) +
theme(strip.background = element_rect(fill="orange2"))
dev.off()
y.lim <- ceiling(max(abs(fit.size.dt.melted[date > as.Date("1980-07-01")]$value)))
jpeg("./inst/etc/img/size_slopes_1980_2018.jpeg", width = 700, height = 500)
ggplot() +
geom_hline(yintercept = 0) +
geom_line(data = fit.size.dt.melted[date > as.Date("1980-7-1")], aes(x = date, y = value), col = "dodgerblue2") +
geom_point(data = fit.size.jan.dt.melted[date > as.Date("1980-7-1")], aes(x = date, y = value), size = 1) +
facet_grid(rows = vars(variable), switch = "both") +
ggtitle("Time Series of Size Slopes 1980-2018") +
xlab("Date") +
scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "5 years", date_minor_breaks = "1 year", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank()) +
theme(strip.background = element_rect(fill="orange2"))
dev.off()
# fit <- lm(beta_1 ~ 1, fit.classic.size.dt[year(date) <= 1990])
#
# size.test <- coeftest(fit, vcov. = NeweyWest(fit, lag = 1, prewhite = FALSE))
standard_error.classic <- sqrt(lrvar(fit.classic.size.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
mu.size.classic <- mean(fit.classic.size.dt[year(date) <= 1990]$beta_1)
t_stat.size.classic <- mu.size.classic / standard_error.classic
standard_error.robust <- sqrt(lrvar(fit.robust.size.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
mu.size.robust <- mean(fit.robust.size.dt[year(date) <= 1990]$beta_1)
t_stat.size.robust <- mu.size.robust / standard_error.robust
size.report.dt <- data.table(Factor = c("", "Size", ""), Method = c("LS (FF92)", "LS (MB)", "Robust (MB)"),
"Slope (t-stat)" = c("-0.15 (-2.58)",
paste0(round(mu.size.classic, 2), " (", round(t_stat.size.classic, 2), ")"),
paste0(round(mu.size.robust, 2), " (", round(t_stat.size.robust, 2), ")")))
jpeg("./inst/etc/img/size_table_ff_comparison.jpeg", width = 700, height = 500)
plot.new()
grid.table(size.report.dt)
dev.off()
## BOOK-TO-MARKET
fit.classic.book_market <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "log_book_market"))
q <- quantile(x$log_book_market, c(0.005, 0.995))
# x$log_book_market <- ifelse(x$log_book_market < q[1], q[1], x$log_book_market)
# x$log_book_market <- ifelse(x$log_book_market > q[2], q[2], x$log_book_market)
x <- x[log_book_market > q[1] & log_book_market < q[2]]
unname(coef(lm(adj_ret * 100 ~ log_book_market, data = x)))
})
fit.robust.book_market <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "log_book_market"))
control <- lmrobdet.control(efficiency = 0.99)
unname(coef(lmrobdetMM(adj_ret * 100 ~ log_book_market, data = x, control = control)))
})
dates <- as.Date(names(fit.classic.book_market))
fit.classic.book_market.dt <- data.table(matrix(unlist(fit.classic.book_market), ncol = 2, byrow = TRUE))
colnames(fit.classic.book_market.dt) <- c("beta_0", "beta_1")
fit.classic.book_market.dt <- cbind(data.table(date = dates), fit.classic.book_market.dt)
fit.robust.book_market.dt <- data.table(matrix(unlist(fit.robust.book_market), ncol = 2, byrow = TRUE))
colnames(fit.robust.book_market.dt) <- c("beta_0", "beta_1")
fit.robust.book_market.dt <- cbind(data.table(date = dates), fit.robust.book_market.dt)
fit.book_market.dt <- data.table(date = dates, LS = fit.classic.book_market.dt$beta_1, ROBUST = fit.robust.book_market.dt$beta_1)
fit.book_market.dt.melted <- melt(fit.book_market.dt, id.vars = "date")
fit.book_market.jan.dt.melted <- fit.book_market.dt.melted[month(date) == 1]
y.lim <- ceiling(max(abs(fit.classic.book_market.dt$beta_1)))
jpeg("./inst/etc/img/LS_book_market_slopes_1963_2018.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.classic.book_market.dt, aes(x = date, y = beta_1), col = "dodgerblue2") +
geom_point(data = fit.classic.book_market.dt[month(date) == 1], aes(x = date, y = beta_1), size = 1) +
geom_hline(yintercept = 0) +
ggtitle("Time Series of LS Book-to-Market Slopes 1963-2018") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank())
dev.off()
y.lim <- ceiling(max(abs(fit.book_market.dt.melted[year(date) <= 1990]$value)))
jpeg("./inst/etc/img/book_market_slopes_1963_1990.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.book_market.dt.melted[year(date) <= 1990], aes(x = date, y = value), col = "dodgerblue2") +
geom_point(data = fit.book_market.jan.dt.melted[year(date) <= 1990], aes(x = date, y = value), size = 1) +
facet_grid(rows = vars(variable), switch = "both") +
geom_hline(yintercept = 0) +
ggtitle("Time Series of Book-to-Market Slopes 1963-1990") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "5 years", date_minor_breaks = "1 year", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank()) +
theme(strip.background = element_rect(fill="orange2"))
dev.off()
y.lim <- ceiling(max(abs(fit.book_market.dt.melted[date > as.Date("1980-7-1")]$value)))
jpeg("./inst/etc/img/book_market_slopes_1980_2018.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.book_market.dt.melted[date > as.Date("1980-7-1")], aes(x = date, y = value), col = "dodgerblue2") +
geom_point(data = fit.book_market.jan.dt.melted[date > as.Date("1980-7-1")], aes(x = date, y = value), size = 1) +
facet_grid(rows = vars(variable), switch = "both") +
geom_hline(yintercept = 0) +
ggtitle("Time Series of Book-to-Market Slopes 1980-2018") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "5 years", date_minor_breaks = "1 year", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank()) +
theme(strip.background = element_rect(fill="orange2"))
dev.off()
# fit <- lm(beta ~ 1, fit.classic.book_market.dt)
#
# book_market.test <- coeftest(fit, vcov. = NeweyWest(fit, lag = 6, prewhite = FALSE))
standard_error.classic <- sqrt(lrvar(fit.classic.book_market.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
mu.book_market.classic <- mean(fit.classic.book_market.dt[year(date) <= 1990]$beta_1)
t_stat.book_market.classic <- mu.book_market.classic / standard_error.classic
standard_error.robust <- sqrt(lrvar(fit.robust.book_market.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
mu.book_market.robust <- mean(fit.robust.book_market.dt[year(date) <= 1990]$beta_1)
t_stat.book_market.robust <- mu.book_market.robust / standard_error.robust
book_market.report.dt <- data.table(Factor = c("", "ln(BE/ME)", ""), Method = c("LS (FF92)", "LS (MB)", "Robust (MB)"),
"Slope (t-stat)" = c("0.50 (5.71)",
paste0(round(mu.book_market.classic, 2), " (", round(t_stat.book_market.classic, 2), ")"),
paste0(round(mu.book_market.robust, 2), " (", round(t_stat.book_market.robust, 2), ")")))
jpeg("./inst/etc/img/book_market_table_ff_comparison.jpeg", width = 700, height = 500)
plot.new()
grid.table(book_market.report.dt)
dev.off()
## Earnings-to-Price
fit.classic.earnings_price <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "earnings_price"))
q <- quantile(x$earnings_price, c(0.005, 0.995))
x$earnings_price <- ifelse(x$earnings_price < q[1], q[1], x$earnings_price)
x$earnings_price <- ifelse(x$earnings_price > q[2], q[2], x$earnings_price)
# x <- x[earnings_price > q[1] & earnings_price < q[2]]
unname(coef(lm(adj_ret * 100 ~ earnings_price, data = x)))
})
fit.classic.earnings_price2 <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "earnings_price"))
q <- quantile(x$earnings_price, c(0.005, 0.995))
# x$earnings_price <- ifelse(x$earnings_price < q[1], q[1], x$earnings_price)
# x$earnings_price <- ifelse(x$earnings_price > q[2], q[2], x$earnings_price)
x <- x[earnings_price > q[1] & earnings_price < q[2]]
unname(coef(lm(adj_ret * 100 ~ earnings_price, data = x)))
})
fit.classic.earnings_price3 <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "earnings_price"))
# q <- quantile(x$earnings_price, c(0.005, 0.995))
# x$earnings_price <- ifelse(x$earnings_price < q[1], q[1], x$earnings_price)
# x$earnings_price <- ifelse(x$earnings_price > q[2], q[2], x$earnings_price)
# x <- x[earnings_price > q[1] & earnings_price < q[2]]
unname(coef(lm(adj_ret * 100 ~ earnings_price, data = x)))
})
fit.robust.earnings_price <- lapply(1:length(data.split), function(i, dt) {
x <- na.omit(data.split[[i]], c("adj_ret", "earnings_price"))
control <- lmrobdet.control(efficiency = 0.99)
if (i %in% c(16, 21, 460)) {
control$family <- "modopt"
}
unname(coef(lmrobdetMM(adj_ret * 100 ~ earnings_price, data = x, control = control)))
},
dt = data.split)
# options(warn=2)
# 16 (1964-10-31), 21 (1965-03-31), 460 (2001-10-31)
dates <- as.Date(names(fit.classic.earnings_price))
# dates[c(16, 21, 460)] "1964-10-31" "1965-03-31" "2001-10-31" <- robust MM not converging on these dates
fit.classic.earnings_price.dt <- data.table(matrix(unlist(fit.classic.earnings_price), ncol = 2, byrow = TRUE))
colnames(fit.classic.earnings_price.dt) <- c("beta_0", "beta_1")
fit.classic.earnings_price.dt <- cbind(data.table(date = dates), fit.classic.earnings_price.dt)
fit.classic.earnings_price.dt2 <- data.table(matrix(unlist(fit.classic.earnings_price2), ncol = 2, byrow = TRUE))
colnames(fit.classic.earnings_price.dt2) <- c("beta_0", "beta_1")
fit.classic.earnings_price.dt2 <- cbind(data.table(date = dates), fit.classic.earnings_price.dt2)
fit.classic.earnings_price.dt3 <- data.table(matrix(unlist(fit.classic.earnings_price3), ncol = 2, byrow = TRUE))
colnames(fit.classic.earnings_price.dt3) <- c("beta_0", "beta_1")
fit.classic.earnings_price.dt3 <- cbind(data.table(date = dates), fit.classic.earnings_price.dt3)
fit.robust.earnings_price.dt <- data.table(matrix(unlist(fit.robust.earnings_price), ncol = 2, byrow = TRUE))
colnames(fit.robust.earnings_price.dt) <- c("beta_0", "beta_1")
fit.robust.earnings_price.dt <- cbind(data.table(date = dates), fit.robust.earnings_price.dt)
fit.earnings_price.dt <- data.table(date = dates, LS = fit.classic.earnings_price.dt3$beta_1,
LS_WINSOR = fit.classic.earnings_price.dt$beta_1, LS_TRIMMED = fit.classic.earnings_price.dt2$beta_1)
fit.earnings_price.dt.melted <- melt(fit.earnings_price.dt, id.vars = "date")
fit.earnings_price.jan.dt.melted <- fit.earnings_price.dt.melted[month(date) == 1]
y.lim <- ceiling(max(abs(fit.classic.earnings_price.dt$beta_1)))
jpeg("./inst/etc/img/LS_earnings_price_slopes_1963_2018.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.classic.earnings_price.dt, aes(x = date, y = beta_1), col = "dodgerblue2") +
geom_hline(yintercept = 0) +
ggtitle("Time Series of Earnings-to-Price Slopes 1963-2018") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank())
dev.off()
y.lim <- ceiling(max(abs(fit.earnings_price.dt.melted[year(date) <= 2015]$value)))
jpeg("./inst/etc/img/ep_chris_1963_2015.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.earnings_price.dt.melted[year(date) <= 2015], aes(x = date, y = value), col = "dodgerblue2") +
facet_grid(rows = vars(variable), switch = "both") +
geom_hline(yintercept = 0) +
ggtitle("Time Series of Earnings-to-Price Slopes 1963-2015") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank()) +
theme(strip.background = element_rect(fill="orange2"))
dev.off()
n_comp.lst <- lapply(data.split, nrow)
dates <- as.Date(ISOdate(year(as.Date(names(n_comp.lst)) %m-% months(6)), 7, 31))
n_comp <- data.table(Start_Date = dates, n = unname(unlist(n_comp.lst)))[year(dates) <= 2015]
n_comp <- n_comp %>%
group_by(Start_Date) %>%
summarise(My_min = min(n),
My_max = max(n)) %>%
as.data.table
chris.n_comp <- fread("/home/gsb25/Desktop/EAPR_additional/data/chris_num_companies.csv")
chris.n_comp[, Start_Date := as.Date(Start_Date)]
n_comp.merged <- merge(chris.n_comp, n_comp, by = "Start_Date")
n_comp.melted <- melt(n_comp.merged, id.vars = "Start_Date")
png("./inst/etc/img/Num_companies.png", width = 800, height = 500)
ggplot() +
geom_point(data = n_comp, mapping = aes(x = Start_Date, y = My_max, color = "Greg", shape = "Max"), size = 1.5) +
geom_point(data = n_comp, mapping = aes(x = Start_Date, y = My_min, color = "Greg", shape = "Min"), size = 1.5) +
geom_point(data = chris.n_comp, mapping = aes(x = Start_Date, y = Chris_max, color = "Chris", shape = "Max"), size = 1.5) +
geom_point(data = chris.n_comp, mapping = aes(x = Start_Date, y = Chris_min, color = "Chris", shape = "Min"), size = 1.5) +
scale_color_manual(values = c("dodgerblue2", "red"))
dev.off()
## Leverage
# fit.classic.leverage <- lapply(data.split, function(x) {
# x <- na.omit(x, c("adj_ret", "log_asset_market", "log_asset_book"))
#
# dat.winsor <- lapply(c("log_asset_market", "log_asset_book"), function(col, dt) {
# q <- quantile(dt[[col]], c(0.005, 0.995))
# temp <- ifelse(dt[[col]] < q[1], q[1], dt[[col]])
# ifelse(temp > q[2], q[2], temp)
# },
# dt = x)
#
# x$log_asset_market <- dat.winsor[[1]]
# x$log_asset_book <- dat.winsor[[2]]
#
# unname(coef(lm(adj_ret * 100 ~ log_asset_market + log_asset_book, data = x)))
# })
#
# fit.robust.leverage <- lapply(data.split, function(x) {
# x <- na.omit(x, c("adj_ret", "log_size"))
# control <- lmrobdet.control(efficiency = 0.99)
# unname(coef(lmrobdetMM(adj_ret * 100 ~ log_asset_market + log_asset_book, data = x, control = control)))
# })
#
# dates <- as.Date(names(fit.classic.leverage))
#
# fit.classic.leverage.dt <- data.table(matrix(unlist(fit.classic.leverage), ncol = 3, byrow = TRUE))
#
# colnames(fit.classic.leverage.dt) <- c("beta_0", "beta_1", "beta_2")
#
# fit.classic.leverage.dt <- cbind(data.table(date = dates), fit.classic.leverage.dt)
#
# fit.robust.leverage.dt <- data.table(matrix(unlist(fit.robust.leverage), ncol = 3, byrow = TRUE))
#
# colnames(fit.robust.leverage.dt) <- c("beta_0", "beta_1", "beta_2")
#
# fit.robust.leverage.dt <- cbind(data.table(date = dates), fit.robust.leverage.dt)
#
# # ASSET-TO-MARKET
#
# fit.leverage.asset_market.dt <- data.table(date = dates, LS = fit.classic.leverage.dt$beta_1, ROBUST = fit.robust.leverage.dt$beta_1)
#
# fit.leverage.asset_market.dt.melted <- melt(fit.leverage.asset_market.dt, id.vars = "date")
#
# fit.leverage.asset_market.jan.dt.melted <- fit.leverage.asset_market.dt.melted[month(date) == 1]
#
# y.lim <- ceiling(max(abs(fit.leverage.asset_market.dt.melted$value)))
#
# ggplot() +
# geom_hline(yintercept = 0) +
# geom_line(data = fit.leverage.asset_market.dt.melted, aes(x = date, y = value), col = "dodgerblue2") +
# geom_point(data = fit.leverage.asset_market.jan.dt.melted, aes(x = date, y = value), size = 1) +
# facet_grid(rows = vars(variable), switch = "both") +
# ggtitle("Time Series of Asset-to-Market Slopes 1963-1990") +
# xlab("Date") +
# scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
# scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
# theme_bw() +
# theme(axis.title.y = element_blank()) +
# theme(strip.background = element_rect(fill="orange2"))
#
# standard_error.classic <- sqrt(lrvar(fit.classic.leverage.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
#
# mu.leverage.asset_market.classic <- mean(fit.classic.leverage.dt[year(date) <= 1990]$beta_1)
#
# t_stat.leverage.asset_market.classic <- mu.leverage.asset_market.classic / standard_error.classic
#
# standard_error.robust <- sqrt(lrvar(fit.robust.leverage.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
#
# mu.leverage.asset_market.robust <- mean(fit.robust.leverage.dt[year(date) <= 1990]$beta_1)
#
# t_stat.leverage.asset_market.robust <- mu.leverage.asset_market.robust / standard_error.robust
#
# leverage.asset_market.report.dt <- data.table(Factor = c("", "ln(A/ME)", ""), Method = c("LS (FF92)", "LS (MB)", "Robust (MB)"),
# "Slope (t-stat)" = c("0.50 (5.69)",
# paste0(round(mu.leverage.asset_market.classic, 2), " (", round(t_stat.leverage.asset_market.classic, 2), ")"),
# paste0(round(mu.leverage.asset_market.robust, 2), " (", round(t_stat.leverage.asset_market.robust, 2), ")")))
#
# plot.new()
# grid.table(leverage.asset_market.report.dt)
#
# # ASSET-TO-BOOK
#
# fit.leverage.asset_book.dt <- data.table(date = dates, LS = fit.classic.leverage.dt$beta_1, ROBUST = fit.robust.leverage.dt$beta_1)
#
# fit.leverage.asset_book.dt.melted <- melt(fit.leverage.asset_book.dt, id.vars = "date")
#
# fit.leverage.asset_book.jan.dt.melted <- fit.leverage.asset_book.dt.melted[month(date) == 1]
#
# y.lim <- ceiling(max(abs(fit.leverage.asset_book.dt.melted$value)))
#
# ggplot() +
# geom_hline(yintercept = 0) +
# geom_line(data = fit.leverage.asset_book.dt.melted, aes(x = date, y = value), col = "dodgerblue2") +
# geom_point(data = fit.leverage.asset_book.jan.dt.melted, aes(x = date, y = value), size = 1) +
# facet_grid(rows = vars(variable), switch = "both") +
# ggtitle("Time Series of Asset-to-Book Slopes 1963-1990") +
# xlab("Date") +
# scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
# scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
# theme_bw() +
# theme(axis.title.y = element_blank()) +
# theme(strip.background = element_rect(fill="orange2"))
#
# standard_error.classic <- sqrt(lrvar(fit.classic.leverage.dt[year(date) <= 1990]$beta_2, type = "Newey-West"))
#
# mu.leverage.asset_book.classic <- mean(fit.classic.leverage.dt[year(date) <= 1990]$beta_2)
#
# t_stat.leverage.asset_book.classic <- mu.leverage.asset_book.classic / standard_error.classic
#
# standard_error.robust <- sqrt(lrvar(fit.robust.leverage.dt[year(date) <= 1990]$beta_2, type = "Newey-West"))
#
# mu.leverage.asset_book.robust <- mean(fit.robust.leverage.dt[year(date) <= 1990]$beta_2)
#
# t_stat.leverage.asset_book.robust <- mu.leverage.asset_book.robust / standard_error.robust
#
# leverage.asset_book.report.dt <- data.table(Factor = c("", "ln(A/BE)", ""), Method = c("LS (FF92)", "LS (MB)", "Robust (MB)"),
# "Slope (t-stat)" = c("-0.57 (-5.34)",
# paste0(round(mu.leverage.asset_book.classic, 2), " (", round(t_stat.leverage.asset_book.classic, 2), ")"),
# paste0(round(mu.leverage.asset_book.robust, 2), " (", round(t_stat.leverage.asset_book.robust, 2), ")")))
#
# plot.new()
# grid.table(leverage.asset_book.report.dt)
GregoryBrownson/EAPR documentation built on Oct. 28, 2019, 7:27 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.
#rm(list = ls())
#library(data.table)
library(devtools)
library(ggplot2)
library(gridExtra)
library(lmtest)
library(purrr)
library(RobStatTM)
library(sandwich)
library(tables)
library(xts)
library(EAPR)
devtools::load_all(".")
eapr <- EAPR::extract("gsb25")
dat.filtered <- EAPR::filter.ff92(eapr)
size.cs <- EAPR::fama_macbeth(adj_ret * 100 ~ log_size, x = dat.filtered)
## SIZE
# LS and Robust regressions
fit.classic.size <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "log_size"))
coef(lm(adj_ret * 100 ~ log_size, data = x))
})
fit.robust.size <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "log_size"))
control <- lmrobdet.control(efficiency = 0.99)
unname(coef(lmrobdetMM(adj_ret * 100 ~ log_size, data = x, control = control)))
})
dates <- as.Date(names(fit.classic.size))
fit.classic.size.dt <- cbind(data.table(date = dates), data.table(Reduce(rbind, fit.classic.size)))
colnames(fit.classic.size.dt) <- c("beta_0", "beta_1")
fit.classic.size.dt <- cbind(data.table(date = dates), fit.classic.size.dt)
fit.robust.size.dt <- data.table(matrix(unlist(fit.robust.size), ncol = 2, byrow = TRUE))
colnames(fit.robust.size.dt) <- c("beta_0", "beta_1")
fit.robust.size.dt <- cbind(data.table(date = dates), fit.robust.size.dt)
fit.size.dt <- data.table(date = dates, LS = fit.classic.size.dt$beta_1, ROBUST = fit.robust.size.dt$beta_1)
fit.size.dt.melted <- melt(fit.size.dt, id.vars = "date")
fit.size.jan.dt.melted <- fit.size.dt.melted[month(date) == 1]
y.lim <- ceiling(max(abs(fit.classic.size.dt$beta_1)))
jpeg("./inst/etc/img/LS_size_slopes_1963_2018.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.classic.size.dt, aes(x = date, y = beta_1), col = "dodgerblue2") +
geom_point(data = fit.classic.size.dt[month(date) == 1], aes(x = date, y = beta_1), size = 1) +
geom_hline(yintercept = 0) +
ggtitle("Time Series of LS Size Slopes 1963-2018") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank())
dev.off()
y.lim <- ceiling(max(abs(fit.size.dt.melted[year(date) <= 1990]$value)))
jpeg("./inst/etc/img/size_slopes_1963_1990.jpeg", width = 700, height = 500)
ggplot() +
geom_hline(yintercept = 0) +
geom_line(data = fit.size.dt.melted[year(date) <= 1990], aes(x = date, y = value), col = "dodgerblue2") +
geom_point(data = fit.size.jan.dt.melted[year(date) <= 1990], aes(x = date, y = value), size = 1) +
facet_grid(rows = vars(variable), switch = "both") +
ggtitle("Time Series of Size Slopes 1963-1990") +
xlab("Date") +
scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "5 years", date_minor_breaks = "1 year", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank()) +
theme(strip.background = element_rect(fill="orange2"))
dev.off()
y.lim <- ceiling(max(abs(fit.size.dt.melted[date > as.Date("1980-07-01")]$value)))
jpeg("./inst/etc/img/size_slopes_1980_2018.jpeg", width = 700, height = 500)
ggplot() +
geom_hline(yintercept = 0) +
geom_line(data = fit.size.dt.melted[date > as.Date("1980-7-1")], aes(x = date, y = value), col = "dodgerblue2") +
geom_point(data = fit.size.jan.dt.melted[date > as.Date("1980-7-1")], aes(x = date, y = value), size = 1) +
facet_grid(rows = vars(variable), switch = "both") +
ggtitle("Time Series of Size Slopes 1980-2018") +
xlab("Date") +
scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "5 years", date_minor_breaks = "1 year", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank()) +
theme(strip.background = element_rect(fill="orange2"))
dev.off()
# fit <- lm(beta_1 ~ 1, fit.classic.size.dt[year(date) <= 1990])
#
# size.test <- coeftest(fit, vcov. = NeweyWest(fit, lag = 1, prewhite = FALSE))
standard_error.classic <- sqrt(lrvar(fit.classic.size.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
mu.size.classic <- mean(fit.classic.size.dt[year(date) <= 1990]$beta_1)
t_stat.size.classic <- mu.size.classic / standard_error.classic
standard_error.robust <- sqrt(lrvar(fit.robust.size.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
mu.size.robust <- mean(fit.robust.size.dt[year(date) <= 1990]$beta_1)
t_stat.size.robust <- mu.size.robust / standard_error.robust
size.report.dt <- data.table(Factor = c("", "Size", ""), Method = c("LS (FF92)", "LS (MB)", "Robust (MB)"),
"Slope (t-stat)" = c("-0.15 (-2.58)",
paste0(round(mu.size.classic, 2), " (", round(t_stat.size.classic, 2), ")"),
paste0(round(mu.size.robust, 2), " (", round(t_stat.size.robust, 2), ")")))
jpeg("./inst/etc/img/size_table_ff_comparison.jpeg", width = 700, height = 500)
plot.new()
grid.table(size.report.dt)
dev.off()
## BOOK-TO-MARKET
fit.classic.book_market <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "log_book_market"))
q <- quantile(x$log_book_market, c(0.005, 0.995))
# x$log_book_market <- ifelse(x$log_book_market < q[1], q[1], x$log_book_market)
# x$log_book_market <- ifelse(x$log_book_market > q[2], q[2], x$log_book_market)
x <- x[log_book_market > q[1] & log_book_market < q[2]]
unname(coef(lm(adj_ret * 100 ~ log_book_market, data = x)))
})
fit.robust.book_market <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "log_book_market"))
control <- lmrobdet.control(efficiency = 0.99)
unname(coef(lmrobdetMM(adj_ret * 100 ~ log_book_market, data = x, control = control)))
})
dates <- as.Date(names(fit.classic.book_market))
fit.classic.book_market.dt <- data.table(matrix(unlist(fit.classic.book_market), ncol = 2, byrow = TRUE))
colnames(fit.classic.book_market.dt) <- c("beta_0", "beta_1")
fit.classic.book_market.dt <- cbind(data.table(date = dates), fit.classic.book_market.dt)
fit.robust.book_market.dt <- data.table(matrix(unlist(fit.robust.book_market), ncol = 2, byrow = TRUE))
colnames(fit.robust.book_market.dt) <- c("beta_0", "beta_1")
fit.robust.book_market.dt <- cbind(data.table(date = dates), fit.robust.book_market.dt)
fit.book_market.dt <- data.table(date = dates, LS = fit.classic.book_market.dt$beta_1, ROBUST = fit.robust.book_market.dt$beta_1)
fit.book_market.dt.melted <- melt(fit.book_market.dt, id.vars = "date")
fit.book_market.jan.dt.melted <- fit.book_market.dt.melted[month(date) == 1]
y.lim <- ceiling(max(abs(fit.classic.book_market.dt$beta_1)))
jpeg("./inst/etc/img/LS_book_market_slopes_1963_2018.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.classic.book_market.dt, aes(x = date, y = beta_1), col = "dodgerblue2") +
geom_point(data = fit.classic.book_market.dt[month(date) == 1], aes(x = date, y = beta_1), size = 1) +
geom_hline(yintercept = 0) +
ggtitle("Time Series of LS Book-to-Market Slopes 1963-2018") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank())
dev.off()
y.lim <- ceiling(max(abs(fit.book_market.dt.melted[year(date) <= 1990]$value)))
jpeg("./inst/etc/img/book_market_slopes_1963_1990.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.book_market.dt.melted[year(date) <= 1990], aes(x = date, y = value), col = "dodgerblue2") +
geom_point(data = fit.book_market.jan.dt.melted[year(date) <= 1990], aes(x = date, y = value), size = 1) +
facet_grid(rows = vars(variable), switch = "both") +
geom_hline(yintercept = 0) +
ggtitle("Time Series of Book-to-Market Slopes 1963-1990") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "5 years", date_minor_breaks = "1 year", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank()) +
theme(strip.background = element_rect(fill="orange2"))
dev.off()
y.lim <- ceiling(max(abs(fit.book_market.dt.melted[date > as.Date("1980-7-1")]$value)))
jpeg("./inst/etc/img/book_market_slopes_1980_2018.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.book_market.dt.melted[date > as.Date("1980-7-1")], aes(x = date, y = value), col = "dodgerblue2") +
geom_point(data = fit.book_market.jan.dt.melted[date > as.Date("1980-7-1")], aes(x = date, y = value), size = 1) +
facet_grid(rows = vars(variable), switch = "both") +
geom_hline(yintercept = 0) +
ggtitle("Time Series of Book-to-Market Slopes 1980-2018") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "5 years", date_minor_breaks = "1 year", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank()) +
theme(strip.background = element_rect(fill="orange2"))
dev.off()
# fit <- lm(beta ~ 1, fit.classic.book_market.dt)
#
# book_market.test <- coeftest(fit, vcov. = NeweyWest(fit, lag = 6, prewhite = FALSE))
standard_error.classic <- sqrt(lrvar(fit.classic.book_market.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
mu.book_market.classic <- mean(fit.classic.book_market.dt[year(date) <= 1990]$beta_1)
t_stat.book_market.classic <- mu.book_market.classic / standard_error.classic
standard_error.robust <- sqrt(lrvar(fit.robust.book_market.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
mu.book_market.robust <- mean(fit.robust.book_market.dt[year(date) <= 1990]$beta_1)
t_stat.book_market.robust <- mu.book_market.robust / standard_error.robust
book_market.report.dt <- data.table(Factor = c("", "ln(BE/ME)", ""), Method = c("LS (FF92)", "LS (MB)", "Robust (MB)"),
"Slope (t-stat)" = c("0.50 (5.71)",
paste0(round(mu.book_market.classic, 2), " (", round(t_stat.book_market.classic, 2), ")"),
paste0(round(mu.book_market.robust, 2), " (", round(t_stat.book_market.robust, 2), ")")))
jpeg("./inst/etc/img/book_market_table_ff_comparison.jpeg", width = 700, height = 500)
plot.new()
grid.table(book_market.report.dt)
dev.off()
## Earnings-to-Price
fit.classic.earnings_price <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "earnings_price"))
q <- quantile(x$earnings_price, c(0.005, 0.995))
x$earnings_price <- ifelse(x$earnings_price < q[1], q[1], x$earnings_price)
x$earnings_price <- ifelse(x$earnings_price > q[2], q[2], x$earnings_price)
# x <- x[earnings_price > q[1] & earnings_price < q[2]]
unname(coef(lm(adj_ret * 100 ~ earnings_price, data = x)))
})
fit.classic.earnings_price2 <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "earnings_price"))
q <- quantile(x$earnings_price, c(0.005, 0.995))
# x$earnings_price <- ifelse(x$earnings_price < q[1], q[1], x$earnings_price)
# x$earnings_price <- ifelse(x$earnings_price > q[2], q[2], x$earnings_price)
x <- x[earnings_price > q[1] & earnings_price < q[2]]
unname(coef(lm(adj_ret * 100 ~ earnings_price, data = x)))
})
fit.classic.earnings_price3 <- lapply(data.split, function(x) {
x <- na.omit(x, c("adj_ret", "earnings_price"))
# q <- quantile(x$earnings_price, c(0.005, 0.995))
# x$earnings_price <- ifelse(x$earnings_price < q[1], q[1], x$earnings_price)
# x$earnings_price <- ifelse(x$earnings_price > q[2], q[2], x$earnings_price)
# x <- x[earnings_price > q[1] & earnings_price < q[2]]
unname(coef(lm(adj_ret * 100 ~ earnings_price, data = x)))
})
fit.robust.earnings_price <- lapply(1:length(data.split), function(i, dt) {
x <- na.omit(data.split[[i]], c("adj_ret", "earnings_price"))
control <- lmrobdet.control(efficiency = 0.99)
if (i %in% c(16, 21, 460)) {
control$family <- "modopt"
}
unname(coef(lmrobdetMM(adj_ret * 100 ~ earnings_price, data = x, control = control)))
},
dt = data.split)
# options(warn=2)
# 16 (1964-10-31), 21 (1965-03-31), 460 (2001-10-31)
dates <- as.Date(names(fit.classic.earnings_price))
# dates[c(16, 21, 460)] "1964-10-31" "1965-03-31" "2001-10-31" <- robust MM not converging on these dates
fit.classic.earnings_price.dt <- data.table(matrix(unlist(fit.classic.earnings_price), ncol = 2, byrow = TRUE))
colnames(fit.classic.earnings_price.dt) <- c("beta_0", "beta_1")
fit.classic.earnings_price.dt <- cbind(data.table(date = dates), fit.classic.earnings_price.dt)
fit.classic.earnings_price.dt2 <- data.table(matrix(unlist(fit.classic.earnings_price2), ncol = 2, byrow = TRUE))
colnames(fit.classic.earnings_price.dt2) <- c("beta_0", "beta_1")
fit.classic.earnings_price.dt2 <- cbind(data.table(date = dates), fit.classic.earnings_price.dt2)
fit.classic.earnings_price.dt3 <- data.table(matrix(unlist(fit.classic.earnings_price3), ncol = 2, byrow = TRUE))
colnames(fit.classic.earnings_price.dt3) <- c("beta_0", "beta_1")
fit.classic.earnings_price.dt3 <- cbind(data.table(date = dates), fit.classic.earnings_price.dt3)
fit.robust.earnings_price.dt <- data.table(matrix(unlist(fit.robust.earnings_price), ncol = 2, byrow = TRUE))
colnames(fit.robust.earnings_price.dt) <- c("beta_0", "beta_1")
fit.robust.earnings_price.dt <- cbind(data.table(date = dates), fit.robust.earnings_price.dt)
fit.earnings_price.dt <- data.table(date = dates, LS = fit.classic.earnings_price.dt3$beta_1,
LS_WINSOR = fit.classic.earnings_price.dt$beta_1, LS_TRIMMED = fit.classic.earnings_price.dt2$beta_1)
fit.earnings_price.dt.melted <- melt(fit.earnings_price.dt, id.vars = "date")
fit.earnings_price.jan.dt.melted <- fit.earnings_price.dt.melted[month(date) == 1]
y.lim <- ceiling(max(abs(fit.classic.earnings_price.dt$beta_1)))
jpeg("./inst/etc/img/LS_earnings_price_slopes_1963_2018.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.classic.earnings_price.dt, aes(x = date, y = beta_1), col = "dodgerblue2") +
geom_hline(yintercept = 0) +
ggtitle("Time Series of Earnings-to-Price Slopes 1963-2018") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank())
dev.off()
y.lim <- ceiling(max(abs(fit.earnings_price.dt.melted[year(date) <= 2015]$value)))
jpeg("./inst/etc/img/ep_chris_1963_2015.jpeg", width = 700, height = 500)
ggplot() +
geom_line(data = fit.earnings_price.dt.melted[year(date) <= 2015], aes(x = date, y = value), col = "dodgerblue2") +
facet_grid(rows = vars(variable), switch = "both") +
geom_hline(yintercept = 0) +
ggtitle("Time Series of Earnings-to-Price Slopes 1963-2015") +
xlab("Date") +
ylab("Slope") +
scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
theme_bw() +
theme(axis.title.y = element_blank()) +
theme(strip.background = element_rect(fill="orange2"))
dev.off()
n_comp.lst <- lapply(data.split, nrow)
dates <- as.Date(ISOdate(year(as.Date(names(n_comp.lst)) %m-% months(6)), 7, 31))
n_comp <- data.table(Start_Date = dates, n = unname(unlist(n_comp.lst)))[year(dates) <= 2015]
n_comp <- n_comp %>%
group_by(Start_Date) %>%
summarise(My_min = min(n),
My_max = max(n)) %>%
as.data.table
chris.n_comp <- fread("/home/gsb25/Desktop/EAPR_additional/data/chris_num_companies.csv")
chris.n_comp[, Start_Date := as.Date(Start_Date)]
n_comp.merged <- merge(chris.n_comp, n_comp, by = "Start_Date")
n_comp.melted <- melt(n_comp.merged, id.vars = "Start_Date")
png("./inst/etc/img/Num_companies.png", width = 800, height = 500)
ggplot() +
geom_point(data = n_comp, mapping = aes(x = Start_Date, y = My_max, color = "Greg", shape = "Max"), size = 1.5) +
geom_point(data = n_comp, mapping = aes(x = Start_Date, y = My_min, color = "Greg", shape = "Min"), size = 1.5) +
geom_point(data = chris.n_comp, mapping = aes(x = Start_Date, y = Chris_max, color = "Chris", shape = "Max"), size = 1.5) +
geom_point(data = chris.n_comp, mapping = aes(x = Start_Date, y = Chris_min, color = "Chris", shape = "Min"), size = 1.5) +
scale_color_manual(values = c("dodgerblue2", "red"))
dev.off()
## Leverage
# fit.classic.leverage <- lapply(data.split, function(x) {
# x <- na.omit(x, c("adj_ret", "log_asset_market", "log_asset_book"))
#
# dat.winsor <- lapply(c("log_asset_market", "log_asset_book"), function(col, dt) {
# q <- quantile(dt[[col]], c(0.005, 0.995))
# temp <- ifelse(dt[[col]] < q[1], q[1], dt[[col]])
# ifelse(temp > q[2], q[2], temp)
# },
# dt = x)
#
# x$log_asset_market <- dat.winsor[[1]]
# x$log_asset_book <- dat.winsor[[2]]
#
# unname(coef(lm(adj_ret * 100 ~ log_asset_market + log_asset_book, data = x)))
# })
#
# fit.robust.leverage <- lapply(data.split, function(x) {
# x <- na.omit(x, c("adj_ret", "log_size"))
# control <- lmrobdet.control(efficiency = 0.99)
# unname(coef(lmrobdetMM(adj_ret * 100 ~ log_asset_market + log_asset_book, data = x, control = control)))
# })
#
# dates <- as.Date(names(fit.classic.leverage))
#
# fit.classic.leverage.dt <- data.table(matrix(unlist(fit.classic.leverage), ncol = 3, byrow = TRUE))
#
# colnames(fit.classic.leverage.dt) <- c("beta_0", "beta_1", "beta_2")
#
# fit.classic.leverage.dt <- cbind(data.table(date = dates), fit.classic.leverage.dt)
#
# fit.robust.leverage.dt <- data.table(matrix(unlist(fit.robust.leverage), ncol = 3, byrow = TRUE))
#
# colnames(fit.robust.leverage.dt) <- c("beta_0", "beta_1", "beta_2")
#
# fit.robust.leverage.dt <- cbind(data.table(date = dates), fit.robust.leverage.dt)
#
# # ASSET-TO-MARKET
#
# fit.leverage.asset_market.dt <- data.table(date = dates, LS = fit.classic.leverage.dt$beta_1, ROBUST = fit.robust.leverage.dt$beta_1)
#
# fit.leverage.asset_market.dt.melted <- melt(fit.leverage.asset_market.dt, id.vars = "date")
#
# fit.leverage.asset_market.jan.dt.melted <- fit.leverage.asset_market.dt.melted[month(date) == 1]
#
# y.lim <- ceiling(max(abs(fit.leverage.asset_market.dt.melted$value)))
#
# ggplot() +
# geom_hline(yintercept = 0) +
# geom_line(data = fit.leverage.asset_market.dt.melted, aes(x = date, y = value), col = "dodgerblue2") +
# geom_point(data = fit.leverage.asset_market.jan.dt.melted, aes(x = date, y = value), size = 1) +
# facet_grid(rows = vars(variable), switch = "both") +
# ggtitle("Time Series of Asset-to-Market Slopes 1963-1990") +
# xlab("Date") +
# scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
# scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
# theme_bw() +
# theme(axis.title.y = element_blank()) +
# theme(strip.background = element_rect(fill="orange2"))
#
# standard_error.classic <- sqrt(lrvar(fit.classic.leverage.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
#
# mu.leverage.asset_market.classic <- mean(fit.classic.leverage.dt[year(date) <= 1990]$beta_1)
#
# t_stat.leverage.asset_market.classic <- mu.leverage.asset_market.classic / standard_error.classic
#
# standard_error.robust <- sqrt(lrvar(fit.robust.leverage.dt[year(date) <= 1990]$beta_1, type = "Newey-West"))
#
# mu.leverage.asset_market.robust <- mean(fit.robust.leverage.dt[year(date) <= 1990]$beta_1)
#
# t_stat.leverage.asset_market.robust <- mu.leverage.asset_market.robust / standard_error.robust
#
# leverage.asset_market.report.dt <- data.table(Factor = c("", "ln(A/ME)", ""), Method = c("LS (FF92)", "LS (MB)", "Robust (MB)"),
# "Slope (t-stat)" = c("0.50 (5.69)",
# paste0(round(mu.leverage.asset_market.classic, 2), " (", round(t_stat.leverage.asset_market.classic, 2), ")"),
# paste0(round(mu.leverage.asset_market.robust, 2), " (", round(t_stat.leverage.asset_market.robust, 2), ")")))
#
# plot.new()
# grid.table(leverage.asset_market.report.dt)
#
# # ASSET-TO-BOOK
#
# fit.leverage.asset_book.dt <- data.table(date = dates, LS = fit.classic.leverage.dt$beta_1, ROBUST = fit.robust.leverage.dt$beta_1)
#
# fit.leverage.asset_book.dt.melted <- melt(fit.leverage.asset_book.dt, id.vars = "date")
#
# fit.leverage.asset_book.jan.dt.melted <- fit.leverage.asset_book.dt.melted[month(date) == 1]
#
# y.lim <- ceiling(max(abs(fit.leverage.asset_book.dt.melted$value)))
#
# ggplot() +
# geom_hline(yintercept = 0) +
# geom_line(data = fit.leverage.asset_book.dt.melted, aes(x = date, y = value), col = "dodgerblue2") +
# geom_point(data = fit.leverage.asset_book.jan.dt.melted, aes(x = date, y = value), size = 1) +
# facet_grid(rows = vars(variable), switch = "both") +
# ggtitle("Time Series of Asset-to-Book Slopes 1963-1990") +
# xlab("Date") +
# scale_y_continuous(position = "right", limits = c(-y.lim, y.lim)) +
# scale_x_date(date_breaks = "10 years", date_minor_breaks = "2 years", date_labels = "%Y") +
# theme_bw() +
# theme(axis.title.y = element_blank()) +
# theme(strip.background = element_rect(fill="orange2"))
#
# standard_error.classic <- sqrt(lrvar(fit.classic.leverage.dt[year(date) <= 1990]$beta_2, type = "Newey-West"))
#
# mu.leverage.asset_book.classic <- mean(fit.classic.leverage.dt[year(date) <= 1990]$beta_2)
#
# t_stat.leverage.asset_book.classic <- mu.leverage.asset_book.classic / standard_error.classic
#
# standard_error.robust <- sqrt(lrvar(fit.robust.leverage.dt[year(date) <= 1990]$beta_2, type = "Newey-West"))
#
# mu.leverage.asset_book.robust <- mean(fit.robust.leverage.dt[year(date) <= 1990]$beta_2)
#
# t_stat.leverage.asset_book.robust <- mu.leverage.asset_book.robust / standard_error.robust
#
# leverage.asset_book.report.dt <- data.table(Factor = c("", "ln(A/BE)", ""), Method = c("LS (FF92)", "LS (MB)", "Robust (MB)"),
# "Slope (t-stat)" = c("-0.57 (-5.34)",
# paste0(round(mu.leverage.asset_book.classic, 2), " (", round(t_stat.leverage.asset_book.classic, 2), ")"),
# paste0(round(mu.leverage.asset_book.robust, 2), " (", round(t_stat.leverage.asset_book.robust, 2), ")")))
#
# plot.new()
# grid.table(leverage.asset_book.report.dt)
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.