Portfolio Risk.R

# Function to calculate portfolio risk
portfolio_risk <- function(x, p.weights, lg = T, sqrt = T){
  
  if (isTRUE(lg)) { x = diff(log(x))[-1,] } # Make logs if needed
  
  # Calculate sum of products for variances and weights
  v.pd<-crossprod(x=as.matrix(apply(x,2,function(col) var(col))),y=p.weights^2)
  
  a.names <- colnames(x) # Assign names
  
  # Create a matrix to store the products of weights
  weight_product_matrix <- matrix(0, nrow=length(a.names),ncol=length(a.names))
  
  # Fill the matrix with weight products
  for (i in 1:length(a.names)) { for (j in 1:length(a.names)) {
      weight_product_matrix[i, j]<-p.weights[i] * p.weights[j] } }
  
  rownames(weight_product_matrix) <- a.names # Row Names
  colnames(weight_product_matrix) <- a.names # Column Names
  
  # Put weights and covariances in nested list
  list_wghts_cov <- list(weight_product_matrix, cov(x))
  
  wghts_cov <- NULL # Create an empty variable to put covariances and weights
  
  for (n in 1:length(list_wghts_cov)){ # For each table in tables
    
    # Extract unique pairs and their correlations
    cor_pairs <- which(upper.tri(list_wghts_cov[[n]], diag = T), arr.ind = T)
    
    # Put them into one data frame
    unique_pairs <- data.frame(
      Variable1 = rownames(list_wghts_cov[[n]])[cor_pairs[, 1]],
      Variable2 = rownames(list_wghts_cov[[n]])[cor_pairs[, 2]],
      Correlation = (list_wghts_cov[[n]])[cor_pairs]
    )
    
    # Filter out pairs where the tickers are not the same
    different_tickers_pairs <- unique_pairs[unique_pairs$Variable1 !=
                                              unique_pairs$Variable2, ]
    
    # Concatenate First_Name and Last_Name with a space in between
    different_tickers_pairs$Pair <- paste(different_tickers_pairs$Variable1,
                                          different_tickers_pairs$Variable2)
    
    # Matrix with values of unique pairs
    new_data_set <- as.data.frame(different_tickers_pairs[,4:3]) 
    
    # Put values into table
    if (is.null(wghts_cov)){ wghts_cov <- new_data_set } else {
      
      wghts_cov_join <- merge(x = wghts_cov, y = new_data_set, by = "Pair")} }
  
  # Sum of products for covariances and weights
  cov.pd <- crossprod(x = wghts_cov_join[,2], y = wghts_cov_join[,3]) * 2
  
  # Portfolio's standard deviation & Portfolio's variance
  if (isTRUE(sqrt)) { p.risk<-(v.pd + cov.pd)^.5 } else { p.risk<-v.pd+cov.pd } 
  
  return(p.risk) # Print the matrix
}
portfolio_risk(x = stock_data, p.weights = c(.4, .3, .2, .1), sqrt = T) # Test