knitr::opts_chunk$set(echo = TRUE,fig.align = "center",
                        message = FALSE,
                      warning = FALSE)
options(tinytex.verbose = TRUE)

1 loading the libraries

library(tidymodels)
library(tidyquant)
library(tidyverse)
library(scales)
library(kableExtra)
library(lubridate)
library(ggplot2)
library(plotly)
library(timetk)
library(modeltime)
library(DataExplorer)
library(correlationfunnel)

#The real estate sector generally includes two different types of investments related to real estate. Some stocks in the sector are responsible for developing new real estate projects and then managing them by obtaining tenants for various spaces within the project property. In addition, most real estate investment trusts, which are special tax-favored business entities that operate in various areas of the real estate industry, get counted as within the real estate sector.

2 Data For real Estate Sector

#tickers
symbols <- c("WELL","WY","LPX","DRE","PEAK") %>% sort()
# prices
# stock prices
prices  <- symbols %>%
    tq_get(get = "stock.prices", from = "1990-01-01")
# returns
returns <- prices %>%
    group_by(symbol) %>%
    tq_transmute(
        select = adjusted,
        mutate_fun = periodReturn,
        period     = "monthly",
    ) %>%
    ungroup() %>%
    mutate(date = rollback(date, roll_to_first = TRUE))
returns

3 Visualization of Returns

returns %>%
    group_by(symbol) %>%
    ggplot(aes(date, monthly.returns, color = symbol)) +
    geom_line() + scale_y_continuous(labels = scales::percent) +
    theme_tq() + facet_wrap(~symbol) + 
    labs(
        title = "Monthly returns",
        subtitle = "Real Estate Sector",
        x = "",
        y = "Monthly Returns"
   )

#timetk

returns %>%
    plot_time_series(
        .date_var = date,
        .value = monthly.returns,
        .facet_vars = symbol,
        .color_var = symbol,
        .facet_ncol = 2,
        .title = "Returns of Real Estate Sector Stocks",
        .x_lab = "",
        .y_lab = "Monthly Returns",
        .smooth_size = 0.5
    )

#Stock Prices Visualization

prices %>%
    group_by(symbol) %>%
    plot_time_series(
        .date_var = date,
        .value = adjusted,
        .color_var = symbol,
        .facet_ncol = 2,
        .facet_scales = "free_y",
        .smooth = FALSE,
        .title = "Stock Prices Real Estate Sector",
        .x_lab = "",
        .y_lab = "Adjusted Prices",
        .legend_show = FALSE
    )

#timetk

prices %>%
    plot_time_series(
        .date_var = date,
        .value = adjusted,
        .facet_vars = symbol,
        .color_var = symbol,
        .facet_ncol = 2,
        .title = "Stock Prices Real Estate Sector",
        .x_lab = "",
        .y_lab = "Adjusted Stock Prices($)",
        .smooth = FALSE
    )

#Portfolio Construction

# Weights Allocation
w <- rep(1/5, 5)
# Portfolio creation
port_ret_tbl <- returns %>%
    tq_portfolio(
        assets_col = symbol, 
        returns_col = monthly.returns,
        weights = w
    ) %>%
    mutate(date = rollback(date, roll_to_first = TRUE)) %>%
    add_column(symbol ="Portfolio", .before = 1)

#Visualization
port_ret_tbl  %>%
    ggplot(aes(date, portfolio.returns)) +
    geom_line() + geom_smooth(method = "loess") +
    scale_y_continuous(labels = scales::percent) + theme_tq() + scale_color_tq() +
    labs(
        title = "Portfolio Returns",
        subtitle = "Real Estate Sector",
        y = "Returns",
        x = ""
    )

#timetk

port_ret_tbl %>%
    plot_time_series(
        .date_var = date,
        .value = portfolio.returns,
        .title = "Portfolio Returns Real Estate Sector",
        .x_lab = "",
        .y_lab = "Returns"
    )

#S&P500

market_prices <- tq_get("^GSPC", get = "stock.prices",from = "1990-01-01")
market_returns <-  market_prices %>%
    tq_transmute(select = adjusted,
                 mutate_fun = periodReturn,
                 period = "monthly",
                 col_rename = "market_ret") %>%
    mutate(date = rollback(date, roll_to_first = TRUE))
market_returns

#FAMA-FRENCH FACTORS

library(frenchdata)
factors_ff_monthly_raw <- download_french_data("Fama/French 3 Factors")
factors_ff_monthly <- factors_ff_monthly_raw$subsets$data[[1]] |>
  transmute(
    date = floor_date(ymd(str_c(date,"01")),"month"),
    rf = as.numeric(RF) / 100,
    mkt_excess = as.numeric(`Mkt-RF`) / 100,
    smb = as.numeric(SMB) / 100,
    hml = as.numeric(HML) / 100
  ) %>%
  # filter the dates 
  filter(date >= "1990-01-01")

factors_ff_monthly

#Combining all the data

port_risk_market_tbl <- port_ret_tbl %>%
    left_join(factors_ff_monthly, by = "date") %>%
    left_join(market_returns, by = "date") %>%
    mutate(excess_returns = portfolio.returns - rf) %>%
    mutate(market_excess_returns = market_ret - rf) %>%
    select(symbol, date, mkt_excess, smb, hml, excess_returns, market_excess_returns)
port_risk_market_tbl

#Beta Estimation ## Step1: CAPM Function

library(slider)
# estimate capm
estimate_capm <- function(data, min_obs = 1) {
  if (nrow(data) < min_obs) {
    beta <- as.numeric(NA)
  } else {
    fit <- lm(excess_returns ~ mkt_excess + smb+hml+ market_excess_returns, data = data)
    beta <- as.numeric(fit$coefficients[2])
  }
  return(beta)
}
# rolling capm estimation
roll_capm_estimation <- function(data, months, min_obs) {
  data <- bind_rows(data) |>
    arrange(date)

  betas <- slide_period_vec(
    .x = data,
    .i = data$date,
    .period = "month",
    .f = ~estimate_capm(., min_obs),
    .before = months - 1,
    .complete = FALSE
  )

  tibble(
    month = unique(data$date),
    beta = betas
  )
}

##Step2: Estimation

beta_port_tbl <- port_risk_market_tbl %>%
    mutate(roll_capm_estimation(cur_data(), months = 60, min_obs = 48)) %>%
    drop_na() %>%
    select(symbol, date, beta)

beta_port_tbl

##Step3: Beta Visualization

beta_port_tbl %>%
    ggplot(aes(date, beta)) +
    geom_line() + geom_smooth(method = "loess") +
    scale_y_continuous() + theme_tq() + scale_color_tq() +
    labs(
        title = "Rolling Beta using 5 years data",
        subtitle = "Real Estate Sector",
        caption = "Stocks(DRE, LPX, PEAK, WELL, WY)",
        x ="",
        y="Beta"
)

#timetk

beta_port_tbl %>%
    plot_time_series(
        .date_var = date,
        .value = beta,
        .title = "Rolling Beta Estimates Real Estate Sector",
        .x_lab = "",
        .y_lab = "Beta"
    )

4 Stock Returns, Fama-French and S&P500

# combining the data
returns_risk_market_tbl <- returns %>%
    group_by(symbol) %>%
    left_join(factors_ff_monthly, by = "date") %>%
    left_join(market_returns, by = "date") %>%
    mutate(excess_returns = monthly.returns -rf) %>%
    mutate(market_excess_returns = market_ret - rf) %>%
    ungroup() %>%
    drop_na() %>%
    select(symbol, date, excess_returns, market_excess_returns, mkt_excess, smb, hml)
returns_risk_market_tbl

##Beta Estimation

beta_stocks_tbl <- returns_risk_market_tbl %>%
    group_by(symbol) %>%
    mutate(roll_capm_estimation(cur_data(), months = 60, min_obs = 48)) %>%
    ungroup() %>%
    drop_na() %>%
    select(symbol, date, beta)
beta_stocks_tbl

##Beta Visualization

beta_stocks_tbl %>%
    ggplot(aes(date, beta, color = symbol)) +
    geom_line() + scale_y_continuous() +
    theme_tq() +
    labs(
        title = "Rolling beta with 5 years of data",
        subtitle = "Real Estate Sector ",
        capition = "stocks(DRE, WELL,WY, LPX, PEAK)",
        y ="Beta",
        x =""
    )

###timetk

beta_stocks_tbl %>%
    plot_time_series(
        .date_var = date,
        .value = beta,
        .color_var = symbol,
        .smooth = FALSE,
        .title = "Rolling Beta for Stocks (Real Estate Sector)",
        .x_lab = "",
        .y_lab = "Beta"
    )