acf() Function plots very large lags

Question

In order to define which Garch Model I should use I tried using the acf function on my daily returns.

I used the following code:

    acf.pax=acf(pax_daily_return,main='ACF PAX',lags.max = 12, ylim=c(- 0.5,1))

However, when I plot it my y axes ranges from -0.4 to 0.8 and my lags range from 0 to 2500000.

Can anyone tell me what I did wrong and how I can fix it?

Answer 1

The X axis is time, not lags. For example, suppose we measure time in seconds and there is an observation every 3 days. Then we have the following where frequency is the reciprocal of deltat. (Try passing ts(pax_daily_return) to acf.)

pax_daily_return <- ts(1:20, deltat = 3 * 24 * 60 * 60)

deltat(pax_daily_return)  # seconds between observations
## [1] 259200

frequency(pax_daily_return)  # samples per second
## [1] 3.858025e-06

acf(pax_daily_return, main = 'ACF PAX', lag.max = 12, ylim = c(-0.5, 1))