topic Two Time Series -- Measure of 'In-Sync' or 'In-Phase' in SAS Forecasting and Econometrics

Two Time Series -- Measure of 'In-Sync' or 'In-Phase'

NKormanik — Wed, 30 Jun 2021 23:28:05 GMT

Suppose we have two time-series, X1 and X2. I'm wondering if SAS has a way of showing if the two series are in-sync, or in-phase, vs. out of sync/phase.

Let's say we want to execute some action when the sync of X1, X2 is maximally moving in the same direction.

We have to be able to track the sync/phase. Have a measure thereof.

Seems what I am seeking is like a 'moving correlation' -- which takes into account the previous so many days of data.

Any thoughts greatly appreciated.

Nicholas Kormanik

Re: Two Time Series -- Measure of 'In-Sync' or 'In-Phase'

KatScott — Thu, 01 Jul 2021 22:41:41 GMT

Nicholas,

Check this useful post:

https://communities.sas.com/t5/New-SAS-User/Moving-Correlation-For-multiple-variables/td-p/559190

Best,

Re: Two Time Series -- Measure of 'In-Sync' or 'In-Phase'

mkeintz — Fri, 02 Jul 2021 03:37:48 GMT

I've always been disappointed that PROC EXPAND only does univariate series transformations. That's why in the useful paper mentioned by @KatScott you have to prepare the dataset (e.g. create product X1_2=x1*x2) prior to PROC EXPAND, which of course is followed by a PROC CORR.

It might be a better idea to skip the proc expand, as in:


data need (drop=_:) view=need ;
  set have;
  if _n_=1 then do;
    declare hash h (dataset:'have(obs=0)',ordered:'a');
      h.definekey('date');
      h.definedata(all:'Y');
      h.definedone();
    declare hiter hi ('h');
  end;

  h.add();
  
  _date&winsize=lag&winsize(date);
  window_close_date=date ;
  format window_close_date date9. ;

  if _date&winsize^=. then h.remove(key:_date&winsize);

  if _n_>=&winsize then do _rc=hi.first() by 0 until (hi.next()^=0);
    output;
  end;
run;


proc corr data=need noprint  out=correlations;
  by window_close_date ;
  var x y;
run;

Dataset view NEED simply creates a set of records for each window. In the above instance the window size is 21, and is identified by the variable WINDOW_CLOSE_DATE. Each WINDOW_SIZE_DATE will have 21 records ( and therefore most records will appear in 21 windows) in NEED. For a large windows size, that's a lot of records, but disk activity is kept to a minimum by generating NEED as a data set VIEW, not a data set FILE.