Given simple returns, how do I generate acf and pacf for for 24 lags?

[lina1583]

hi all,

i'm very new to SAS but have programming experience with java. just wondering if anyone has an example code snippet that reads simple return from a file and compute the acf and pacf for 24 lags? thank you for your help in advance.

data will be in this format

date                         r

19700130              0.05

19700227              0.06

19700331              -0.03

......

.......

......

[PGStats]

The identify statement in proc arima produces acf and pacf estimates. Proc arima is part of SAS/ETS.

[lina1583]

hi, thanks for your reply. what are the statements that I need to write down please? I'm quite new to SAS so i'm not even sure how i should write the identify statements.

so if i have the below, what needs to be changed to output acf and pacf?

data crsp;

infile "\\C\\test.txt"

input date dec1;

proc arima data=crsp;

identify var=dec1 nlag=24 stationarity=(adf);

run;

[PGStats]

See introductory example in the proc ARIMA documentation

https://documentation.sas.com/?docsetId=etsug&docsetTarget=etsug_arima_gettingstarted02.htm&docsetVe...

[lina1583]

thanks, i can see the acf and pacf's graph are there but I need a table which has all the numbers for ACF and PACF like below. how do I do it?

and how do I run Ljung-box test to test my hypothesis that first 12 lags ACF is 0?

lag   acf    pacf

1       0.2     0.2

2       0.1     0.1

3       0.05    0.05

...

23

24

[PGStats]

To get the graph data you will need to use ods output statement:

ODS OUTPUT AutoCorrGraph=myACFdata PACFgraph=myPACFdata;

You should be getting the Ljung-box statistic in a table out of the identify statement output. Specify whitenoise=ignoremiss if your data contains missing values.