Calculate period return from monthly returns - r

Question

This may sound naive but I can't seem to find the solution. I need to calculate 1, 3 and 5-year returns and my dataset consists of monthly returns rather than prices. The dataset I'm working on is similar to managers
data(managers)
tail(managers)
HAM1 HAM2 HAM3 HAM4 HAM5 HAM6 EDHEC LS EQ SP500 TR US 10Y TR US 3m TR
2006-07-31 -0.0144 -0.0131 0.0102 -0.0120 -0.0164 -0.0225 -0.0031 0.00620 0.01580 0.00423
2006-08-31 0.0161 -0.0113 0.0253 -0.0183 0.0169 0.0193 0.0114 0.02380 0.02190 0.00441
2006-09-30 0.0068 -0.0231 0.0072 0.0197 0.0132 -0.0177 0.0001 0.02580 0.01140 0.00456
2006-10-31 0.0427 0.0167 0.0183 0.0518 0.0266 0.0189 0.0194 0.03260 0.00584 0.00381
2006-11-30 0.0117 0.0206 0.0269 0.0373 0.0038 0.0300 0.0200 0.01900 0.01419 0.00430
2006-12-31 0.0115 -0.0062 0.0110 0.0206 0.0317 0.0215 0.0153 0.01403 -0.01550 0.00441
I looked into the Return.cumulative from package PerformanceAnalytics but there is no argument for specifying periods. ROC from TTR can specify the number of periods to use but it is not based on return. What would be the best way to do this? Thank you in advance!

Answer 1

Base on what you want and what you know about ROC from TTR , I will only provide the Data preparation part
#Sample Data
df=read.table(text=' Date HAM1 HAM2 HAM3 HAM4 HAM5 HAM6
2006-07-31 -0.0144 -0.0131 0.0102 -0.0120 -0.0164 -0.0225
2006-08-31 0.0161 -0.0113 0.0253 -0.0183 0.0169 0.0193
2006-09-30 0.0068 -0.0231 0.0072 0.0197 0.0132 -0.0177
2006-10-31 0.0427 0.0167 0.0183 0.0518 0.0266 0.0189
2006-11-30 0.0117 0.0206 0.0269 0.0373 0.0038 0.0300
2006-12-31 0.0115 -0.0062 0.0110 0.0206 0.0317 0.0215 ',header=T,stringsAsFactors=F)
#Make the Return to Price assume all stocks initial value with 100
for (i in 2:dim(df)[2]){
B=Reduce(function(x,y) {x * (1+y)}, df[,i], init=100, accumulate=T)# if it is log Return: {x * exp(y)}
if (i==2){
Price= B
}else{
Price=cbind(Price,B)
}
}
Price=data.frame(cbind(df$Date,Price[-1,]))
names(Price)=names(df)
> Price
Date HAM1 HAM2 HAM3 HAM4 HAM5 HAM6
1 2006-07-31 98.56 98.69 101.02 98.8 98.36 97.75
2 2006-08-31 100.146816 97.574803 103.575806 96.99196 100.022284 99.636575
3 2006-09-30 100.8278143488 95.3208250507 104.3215518032 98.902701612 101.3425781488 97.8730076225
4 2006-10-31 105.133162021494 96.9126828290467 106.230636201199 104.025861555502 104.038290727558 99.7228074665652
5 2006-11-30 106.363220017145 98.909084095325 109.088240315011 107.906026191522 104.433636232323 102.714491690562
6 2006-12-31 107.586397047342 98.295847773934 110.288210958476 110.128890331067 107.744182500887 104.922853261909
Then you can use the normal package to annualized the return(or customize)

Answer 2

Using #Wen's data:
df=read.table(text=' Date HAM1 HAM2 HAM3 HAM4 HAM5 HAM6
2006-07-31 -0.0144 -0.0131 0.0102 -0.0120 -0.0164 -0.0225
2006-08-31 0.0161 -0.0113 0.0253 -0.0183 0.0169 0.0193
2006-09-30 0.0068 -0.0231 0.0072 0.0197 0.0132 -0.0177
2006-10-31 0.0427 0.0167 0.0183 0.0518 0.0266 0.0189
2006-11-30 0.0117 0.0206 0.0269 0.0373 0.0038 0.0300
2006-12-31 0.0115 -0.0062 0.0110 0.0206 0.0317 0.0215 ',header=T,stringsAsFactors=F)
You can use the rollaply function from the zoo package.
library(zoo)
roll <- function(x, n, stat) {
if (length(x) <= n) NA
else x <- x + 1
rollapply(x, list(-seq(n)), stat, fill = NA)
}
df2 <- transform(df, four_month_return_HAM1 = ave(HAM1, FUN = function(x) roll(x, 4, prod)-1))
Change 4 to the period you want to calculate the cumulative return over. So, for one year, this would be 12. This will then give you the 12 month rolling returns.