Seasonal Arima

[jag_211085]

I want to estimate an ARIMA(2,0,0)(0,0,2)[4] model with a co-efficient on my variable called fee.

I put the 4 in the square brackets there to show it is quarterly data.

Doing:

eq x.ls fee c ar(1) ar(2) sma(4)

gives me output with the ar terms but an ma term instead of sma.

Can you please suggest what I am doing wrong.

Thanks.

[trubador]

You are not doing anything wrong, it is just the fact that SMA or SAR is a multiplicative term and that it will work only if there is a related MA or AR term. These terms allow you to from the product of lag polynomials and save you additional degrees-of-freedom. It is generally advisable not to use a mix of AR and MA terms, unless you have strong econometric evidence. So, you can estimate either of the following equations:

Code: Select all

eq x.ls fee c ar(1) ar(2) sar(4) eq x.ls fee c ma(1) ma(2) sma(4)

which are respectively the more compact form of:

Code: Select all

eq x.ls fee c ar(1) ar(2) ar(4) ar(5) ar(6) eq x.ls fee c ma(1) ma(2) ma(4) ma(5) ma(6)