topic Re: Regression model for a time series data set in New SAS User

Regression model for a time series data set

lauticisterna — Thu, 14 Nov 2019 17:01:45 GMT

Hi, thanks for bringing me this space to ask my doubts.

I want to make a regression model to explain one variable with 4 other macroeconomic variables.

I have month period data from 2011 to 2018 of the variable I want to explain, lets suppose Y.

And then I have four explanatory variables to explain the behaviour of the Y variable. Suppose Z, X, C and V.

I was thinking to use a VAR model, with the proc varmax model but I do not really know how it works on SAS.

If someone could help me it would be great.

Thank you so much!

Re: Regression model for a time series data set

Reeza — Thu, 14 Nov 2019 18:27:09 GMT

Have you checked out the examples in the documentation?

Specific example

https://documentation.sas.com/api/docsets/etsug/15.1/content/etsug_code_varex01.htm?locale=en

List of all examples

https://documentation.sas.com/?docsetId=etsug&docsetVersion=15.1&docsetTarget=etsug_varmax_examples.htm&locale=en

They seem to be exactly what you're trying to do.

@lauticisterna wrote:

Hi, thanks for bringing me this space to ask my doubts.

I want to make a regression model to explain one variable with 4 other macroeconomic variables.

I have month period data from 2011 to 2018 of the variable I want to explain, lets suppose Y.

And then I have four explanatory variables to explain the behaviour of the Y variable. Suppose Z, X, C and V.

I was thinking to use a VAR model, with the proc varmax model but I do not really know how it works on SAS.

If someone could help me it would be great.

Thank you so much!

Re: Regression model for a time series data set

lauticisterna — Thu, 14 Nov 2019 20:07:29 GMT

Hello thank you so much for your answers!

I have already checked those examples. The examples are pretty good for illustrative and operational objectives. The problem is that there is no explanation about the lags selection, or any statistical significance. Also the example of US does not show the p-value for the Dickey-Fuller test for testing stationarity.

But I think that maybe I can handle with those examples.

Anyway, thank you so much for the quick respond!

Re: Regression model for a time series data set

Reeza — Thu, 14 Nov 2019 20:10:05 GMT

That's because those are not typically part of the regression exactly, they're part of the initial analysis ahead of time and belong to other procs. PROC AUTOREG and TIMESERIES can help with those. See their examples.

Re: Regression model for a time series data set

lauticisterna — Fri, 15 Nov 2019 02:39:32 GMT

Thank you Reeza! One last question, Do you know an specific proc for find the lags for a var model? I couldn't find it in the proc time series and autores.

Autoreg was great for granger causality.

Thanks again and sorry for bothering!

Re: Regression model for a time series data set

Reeza — Fri, 15 Nov 2019 03:14:24 GMT

Maybe this: https://communities.sas.com/t5/SAS-Forecasting-and-Econometrics/Using-Cross-Correlation-to-determine-time-delay-between-two/td-p/441481