CODEWITHSUNDEEP
rstatisticsregressionquantitative-financerolling-computation
huge
huge

Reputation: 31

Rolling window to calculate Value at Risk in R

I'm sure this is very obvious but i'm a begginer in R and i spent a good part of the afternoon trying to solve this...

I'm trying to create a rolling window to calculate the Value at Risk (VaR) over time.

I already calculated the unconditional VaR for my entire timeserie of 7298 daily returns.

Now, what i'm trying to do is do a rolling window that calculates VaR for a window of 25 days that will roll every one observation for my entire timeserie.

I tried

apply.rolling(nas, trim = TRUE, gap = 25, by = 1, FUN = function(x) VaR(R = nas, p = 0.99, method="historical"))

and

rollapply(nas, width = 25, FUN = function(x) VaR(R = nas, p = 0.99, method="historical"))

where nas is my time serie.

My code still runs from an hour ago... I don't know what I did wrong...

Thank you very much in advance for any help you can provide.

H.

Upvotes: 1

Views: 1474

Answers (1)

Ben Toh
Ben Toh

Reputation: 782

It should be:

rollapply(nas, width = 25, FUN = function(x) VaR(R = x, p = 0.99, method="historical"))

Basically you applying a function that takes in value x (which is a filtered nas to 25 time units), and produces output based on the x. In your original attempt, the function was function(x) VaR(R = nas, p = 0.99, method="historical"), so it takes in value x, but still calculate VaR of the whole nas, and it does that >7000 times, hence it takes forever.

Upvotes: 2

Related Questions