Forecasting out-of-sample with lagged EXPLANATORY variables

[meganmao]

I'm trying to forecast out-of-sample for my dependent variable using my exogeneous variables.

I don't have data for my dependent variable for the forecasting period, but I do my exogeneous are available for the forecasting period. Currently my dependent variable is NA for the forecasting period.

When I have the Forecast box up, I have the following options:

Forecast Equation: 'Ignore Formulae within Series' or 'Substitute Formulae within Series'

Series to Forecast:
PRODDAYS or PRODDAYS/PRODDAYS_SF

The following is my estimation equation:

• Estimation Equation:
  ============================
  PRODDAYS/PRODDAYS_SF = C(1) + C(2)*TREND + C(3)*MARGIN(-1)
  
  Substituted Series Formula
  ============================
  (PROD/DAYS)/PRODDAYS_SF = C(1) + C(2)*TREND + C(3)*(PRICE(-1)-0.5*COST1(-1)

I'm not sure what I should put as my Forecast sample and if I should check Insert actuals for out-of-sample observations

What should I specify in my Forecast options given my estimation equation?

Let's say my sample is from 2006m01 2013m06. I'm trying to forecast for 2013m07 to 2014m06.

Any help is appreciated!

[EViews Gareth]

Enter:

Code: Select all

2013m07 2014m06

in the forecast sample box.

Checking or not checking Insert Actuals will not affect the forecasted values.

[meganmao]

Thanks for your reply.

When I do that, it says 'Unable to compute due to missing data'.

Does this have anything to do with my seasonality factor (which is not available in the forecasting period) or lagged explanatory variable?

[EViews Gareth]

Yes.

[meganmao]

Am I not able to forecast or how do I overcome this?

[EViews Gareth]

If your dependent variable depends on other variables, then you will be unable to forecast that variable without having data for the explanatory variables.

[meganmao]

Got it. I ended up just copying the previous year's seasonality factor for the forecast period and it worked.

Thanks!!