Endogenous & Exogenous Variables

[brennan6738]

Can someone shed some light for me as to the difference between endogenous and exogenous variables when estimating a VAR? I am trying to forecast treasury rates, and I want to forecast the rates using GDP, Inflation, the Dollar, etc. Does this mean that I enter the rate I want to forecast as endogenous and my independent variables (GDP, Inflation, etc.) as exogenous? If I do that it only seems to show one coefficent for ech independent variable when I would think I want to capture the prior relationships of my independent variables and the rate I wish to forercast. Anyone have a suggestion?

[startz]

Remember that to forecast you need the future values of the variables. The endogenous variables are forecast by the var. The exogenous variables you have to supply future values for. (Note, this is the var forecasting interpretation. The meanings are different in structural estimation.)

[hum2004]

hi,
if you estimate a VAR, use block exogenity test / granger causality. If you are working with a VECM (which I asume you do as i suppose that rates are I(1), test the significance of the loading coefficients. Without knowing your data / model, i would suggest that GDP is exogenous, as it most likely is I(0).