Calculate time-weighted returns for a time series

Question

In R, I have an xts object with monthly returns, which I am looking to calculate an overall time-weighted return for each asset individually. Here is a sample of the data:

                SPY      EFA   
 2005-02-28   0.0206   0.0371  
 2005-03-31  -0.0184  -0.0265  
 2005-04-29  -0.0189  -0.0163  

For example, I'm looking to calculate the time-weighted returns for SPY from 2/28/05 through 4/29/05. Manually, I would calculate as [(1 + .0206)*(1 + -.0184) * (1 + .0189) - 1] * 100. I have 100 vectors of assets. How would I accomplish this in R? Thank you.

Answer 1

You can choose sapply and prod with anonymous function to calculate it, like below:

df <- data.frame( spy = c(.0206,-0.0184,0.0189 ), efa = c(0.0371,-0.0265,-0.01631))

sapply(df,function(x)(prod(x+1)-1)*100)

Output:

> sapply(df,function(x)(prod(x+1)-1)*100)
       spy        efa 
 2.0755376 -0.6850001 

Manually multiplying your expression gives:

((1 + .0206)*(1 + -.0184) * (1 + .0189) - 1) * 100 =  2.075538 (approx)

which is same as sapply result for spy