Autoregressive model with exogenous covariates ?

[Matiou]

Hi,

I would like to model a time series where the value depends on its lag, and also on two covariates. More precisely, this is the last equation in the following picture :

http://img297.imageshack.us/img297/2862/eqwh5.jpg

just for your information, LE is Life Expectancy, EXP is Expenditure, and INV is Innovation.

And the problem is, I just don't how to model it with SAS, let alone interpret it. I know AR(1) models, but I've never learnt what to do when you put some other variables in. I found some articles about this, but nothing explaining how to do it with SAS, which is why I seek for help here.

Thanks in advance

Message was edited by: Matiou Message was edited by: Matiou

[Doudou]

2 suggestions.

1. When you use PROC ARIMA, you can specify CROSSCOR=variable-list in the INDENTIFY statement and INPUT=variable-list in ESTIMATE statement.

2. When you use PROC UCM, you can specify dep_variable=variable-list in MODEL statement and LAGS=n in DEPLAG statement. Message was edited by: Doudou