Hello there,
I am currently working on a forecasting model for time charter rates in the shipping market. The last step is to forecast future periods, i.e. the next 8 or so quarters. My data set contains many variables, with time ranging from 2000q1 until 2014q1. The dependent variable TC_rates is regressed on 3 lags of TC_rates, then on a couple of lags of 3 exogenous variables, a couple of multiplicative dummy variables due to a structural break I've found and a lag of a error correction term variable.
Now what I did is this:
arima d.lTC_2750_612_01 d.l1.lTC_2750_612_01 d.l2.lTC_2750_612_01 ............. break_tradel1 if tin(, 2014q1), hessian
predict TChat, y
predict TChatdy, dynamic(tq(2014q1)) y
tsline TChat TChatdy lTC_2750_612_01
This should give me the forecasted values of the time charter rates until the end of my time variable, i.e. 2016q1. Btw, for all my explanatory variables (not the lagged dependent variables of course) I have entered forecasted values in the data set. But when I do all this, I'll only get a forecast of the dependent variable for one quarter, so only for 2014q2. When I do the same with a standard ARIMA model I get values until 2016q1 no problem. Why is it not working? Does anybody have a suggestion?
Thanks!
I am currently working on a forecasting model for time charter rates in the shipping market. The last step is to forecast future periods, i.e. the next 8 or so quarters. My data set contains many variables, with time ranging from 2000q1 until 2014q1. The dependent variable TC_rates is regressed on 3 lags of TC_rates, then on a couple of lags of 3 exogenous variables, a couple of multiplicative dummy variables due to a structural break I've found and a lag of a error correction term variable.
Now what I did is this:
arima d.lTC_2750_612_01 d.l1.lTC_2750_612_01 d.l2.lTC_2750_612_01 ............. break_tradel1 if tin(, 2014q1), hessian
predict TChat, y
predict TChatdy, dynamic(tq(2014q1)) y
tsline TChat TChatdy lTC_2750_612_01
This should give me the forecasted values of the time charter rates until the end of my time variable, i.e. 2016q1. Btw, for all my explanatory variables (not the lagged dependent variables of course) I have entered forecasted values in the data set. But when I do all this, I'll only get a forecast of the dependent variable for one quarter, so only for 2014q2. When I do the same with a standard ARIMA model I get values until 2016q1 no problem. Why is it not working? Does anybody have a suggestion?
Thanks!
